Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.

Settings

Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup

Settings


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Twitter Facebook YouTube Mastodon MeWe

RSS Posts RSS Comments Email Subscribe


Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...



Username
Password
New? Register here
Forgot your password?

Latest Posts

Archives

On the reliability of the U.S. Surface Temperature Record

Posted on 22 January 2010 by John Cook

The website surfacestations.org enlisted an army of volunteers, travelling across the U.S. photographing weather stations. The point of this effort was to document cases of microsite influence - weather stations located near car parks, air conditioners and airport tarmacs and anything else that might impose a warming bias. While photos can be compelling, the only way to quantify any microsite influence is through analysis of the data. This has been done in On the reliability of the U.S. Surface Temperature Record (Menne 2010), published in the Journal of Geophysical Research. The trends from poorly sited weather stations are compared to well-sited stations. The results indicate that yes, there is a bias associated with poor exposure sites. However, the bias is not what you expect.

Weather stations are split into two categories: good (rating 1 or 2) and bad (ratings 3, 4 or 5). Each day, the minimum and maximum temperature are recorded. All temperature data goes through a process of homogenisation, removing non-climatic influences such as relocation of the weather station or change in the Time of Observation. In this analysis, both the raw, unadjusted data and homogenised, adjusted data are compared. Figure 1 shows the comparison of unadjusted temperature from the good and bad sites. The top figure (c) is the maximum temperature, the bottom figure (d) is the minimum temperature. The black line represents well sited weather stations with the red line representing poorly sited stations.

Maximum and Minimum Temperature Anomaly for good and bad sites
Figure 1. Annual average maximum and minimum unadjusted temperature change calculated using (c) maximum and (d) minimum temperatures from good and poor exposure sites (Menne 2010).

Poor sites show a cooler maximum temperature compared to good sites. For minimum temperature, the poor sites are slightly warmer. The net effect is a cool bias in poorly sited stations. Considering all the air-conditioners, BBQs, car parks and tarmacs, this result is somewhat a surprise. Why are poor sites showing a cooler trend than good sites?

The cool bias occurs primarily during the mid and late 1980s. Over this period, about 60% of USHCN sites converted from Cotton Region Shelters (CRS otherwise known as Stevenson Screens) to electronic Maximum/Minimum Temperature Systems (MMTS). MMTS sensors are attached by cable to an indoor readout device. Consequently, limited by cable length, they're often located closer to heated buildings, paved surfaces and other artificial sources of heat.

Investigations into the impact of the MMTS on temperature data have found that on average, MMTS sensors record lower daily maximums than their CRS counterparts, and, conversely, slightly higher daily minimums (Menne 2009). Only about 30% of the good sites currently have the newer MMTS-type sensors compared to about 75% of the poor exposure locations. Thus it's MMTS sensors that are responsible for the cool bias imposed on poor sites.

When the change from CRS to MMTS are taken into account, as well as other biases such as station relocation and Time of Observation, the trend from good sites show close agreement with poor sites.

Maximum and Minimum Temperature Anomaly for good and bad sites
Figure 2: Comparison of U.S. average annual (a) maximum and (b) minimum temperatures calculated using USHCN version 2 adjusted temperatures. Good and poor site ratings are based on surfacestations.org.

Does this latest analysis mean all the work at surfacestations.org has been a waste of time? On the contrary, the laborious task of rating each individual weather station enabled Menne 2010 to identify a cool bias in poor sites and isolate the cause. The role of surfacestations.org is recognised in the paper's acknowledgements in which they "wish to thank Anthony Watts and the many volunteers at surfacestations.org for their considerable efforts in documenting the current site characteristics of USHCN stations." A net cooling bias was perhaps not the result the surfacestations.org volunteers were hoping for but improving the quality of the surface temperature record is surely a result we should all appreciate.

UPDATE 24/1/2010: There seems to be some confusion in the comments mistaking Urban Heat Island and microsite influences which are two separate phenomenon. Urban Heat Island is the phenomenon where a metropolitan area in general is warmer than surrounding rural areas. This is a real phenomenon (see here for a discussion of how UHI affects warming trends). Microsite influences refer to the configuration of a specific weather station - whether there are any surrounding features that might impose a non-climatic bias.

UPDATE 24/1/2010: There has been no direct response from Anthony Watts re Menne 2010. However, there was one post yesterday featuring a photo of a weather station positioned near an air-conditioner along with the data series from that particular station showing a jump in temperature. The conclusion: "Who says pictures don’t matter?"

So the sequence of events is this. Surfacestations.org publishes photos and anecdotal evidence that microsite influences inflate the warming trend but no data analysis to determine whether there's any actual effect on the overall temperature record. Menne 2010 performs data analysis to determine whether there is a warming bias in poorly position weather stations and finds overall, there is actually a cooling bias. Watts responds with another photo and single piece of anecdotal evidence.

UPDATE 28/1/2010: Anthony Watts has posted a more direct response to Menne 2010 although he admits it's not complete, presumably keeping his powder dry for a more comprehensive peer reviewed response which we all eagerly anticipate. What does this response contain?

More photos, for starters. You can never have enough photos of dodgy weather stations. He then rehashes an old critique of a previous NOAA analysis criticising the use of homogenisation of data. This is curious considering Menne 2010 makes a point of using unadjusted, raw data and in fact, it is this data that reveals the cooling bias. I'm guessing he was so enamoured with the water pollution graphics, he couldn't resist reusing them (the man does recognise the persuasive power of a strong graphic).

0 0

Printable Version  |  Link to this page

Comments

1  2  3  4  Next

Comments 1 to 50 out of 177:

  1. Oh dear, Watts will not be a happy camper. I wonder how long until we see this as a headline on WUWT? I'm not holding my breath. Point is, after reading about a dozen different articles on the issue, I've known all along that the so-called Urban Heat Island effect was nothing more than an Urban Legend!
    0 0
  2. Oh dear, poor deniers. Another of their imaginary supportive planks falls away beneath them...
    0 0
  3. A clear and unambiguous analysis. However, I think it is optimistic to think the outcome of this survey will be presented by Watts or other skeptic bloggers in the way it has been here. If they mention it at all, it will be in a way that shows (if not proves!) somehow or other that the results they were hoping for have indeed been found. Several studies published in SCIENCE a few years ago pretty much demolished the notion of the urban heat island, but since when have such things ever troubled the denialist/skeptic/contrarian campaign? Admittedly, since the field research was organised by their own sympathisers in this occasion, it will be harder to conclude that corrupted scientists and the socialist/liberal/big government conspiracy has rigged the data. But never underestimate the inventiveness of the paranoid mind.
    0 0
  4. Watts writes that "the GISS data isn’t much to be trusted," but he doesn't say why.
    0 0
  5. My guess is that it wasn't deniers that instigated the rating process of weather stations but good old climatologist way back. What was the initial justification for this? And Marcus again it wasn't deniers who invented the Urban Heat Island idea. I just put this into Google Scholar search with many limitations and got over 400,000 hits (900,000 without the limitations). As far as I'm aware denier websites don't show in scholar searches.
    0 0
  6. http://ams.confex.com/ams/15AppClimate/techprogram/paper_91613.htm "The National Weather Service MMTS (Maximum-Minimum Temperature System) -- 20 years after Nolan J. Doesken, Colorado State Univ., Ft. Collins, CO During the mid 1980s, the National Weather Service began deploying electronic temperature measurement devices as a part of their Cooperative Network. The introduction of this new measurement system known as the MMTS (Maximum-Minimum Temperature System) represented the single largest change in how temperatures were measured and reported since the Cooperative Network was established in the 1800s. Early comparisons of MMTS readings with temperature measurements from the traditional liquid-in-glass thermometers mounted in Cotton Region shelters showed small but significant differences. During the first decade, several studies were conducted and published results showed that maximum temperatures from the MMTS were typically cooler and minimum temperatures warmer compared to traditional readings. This was a very important finding affecting climate data continuity and the monitoring of local, regional and national temperature trends. It has now been 20 years since the initial deployment of the MMTS. The Colorado Climate Center at Colorado State University has continued side by side daily measurements with both the MMTS and the traditional liquid-in-glass thermometers. This paper presents a 20-year comparison of temperatures measured 4 meters apart. Results show that little has changed in the relationship between MMTS and liquid-in-glass. Despite a yellowing of the MMTS radiation shield over time, the MMTS continues to read cooler during the daylight hours at all times of year. Minimum temperatures show little difference but with a small seasonal cycle in temperature differences. The largest differences continue, as they were first observed in 1985, to occur with low sun angles, clear skies, light winds and fresh snowcover. In addition to quantitative comparisons, some general comments on the impact of MMTS and other electronic temperature measurements on long-term temperature measurements and observed trends will also be offered." So it is consistent with the findings in 2005
    0 0
  7. Typo in 3 places: "Memme 2010" should be "Menne 2010". So the Watts project shows there is, if anything, an overall cool bias in the raw data, which is a moot point considering the adjustments remove most of this bias. "Adjustments applied to USHCN Version 2 data largely account for the impact of instrument and siting changes, although a small overall residual negative (“cool”) bias appears to remain in the adjusted maximum temperature series." One criticism of the Menne result is that they are relying on the data the Watts army of volunteers puts together with regards to the rating classification. How reliable is that data? Do they have the expertise and objectivity needed to effectively assign a rating to each station? The Watts project has served it's purpose, which is to spread doubt about the data among the laypersons. Peer-reviewed academic studies will just be dismissed as being part of the hoax. How can they be believed, when "photos" prove otherwise?
    0 0
    Response: Thanks for the alert, I've fixed the typo. Re the issue of relying on the surfacestations.org classifications, the NOAA also have their own independent ratings. The dotted lines in the figures above represent the good/bad trends according to their own classifications while the solid lines are according to the surfacestations.org classifications.
  8. It's one thing to endlessly kvetch and complain, another to actually exert some effort. Watts and crew did some work, hats off to them, misguided though they were. I hope these results will encourage doubters to purchase historical weather records for those locations where they are complaining about interpolation. As NewYorkJ remarks, many doubters will play the "hoax" wildcard yet again in order to explain this latest disaster for their strange cause. However, each time the doubt community must draw that card from their hand the remaining slice of the behavioral bell curve containing them loses area.
    0 0
  9. I don't think the UHI can be regarded as a myth. I can't comment on the US and haven't yet read the Memme/Menne paper. Perhaps the UHI is neglible in the US as Peterson found. But a 2009 IJC paper: Detection of urban warming in recent temperature trends in Japan, by Fumiaki Fujibe, showed a 0.1'C/decade UHI effect for the larger cities. This was based on 1979-2006 from 561 stations recording hourly data and compared with local population density data. You can see more about this paper at http://scienceofdoom.com/2010/01/17/urban-heat-island-in-japan/ There is also the Ren paper from 2008 in Journal of Climate which also found a significant UHI in China.
    0 0
  10. Stevecarsonr you're confusing UHI and microsite issues. Nobody believes that the UHI is a myth and there is an abundant litterature on the subject. GISTEMP corrects for UHI. Many papers about it were mentioned on this very blog. Watts' basic argument, insofar as it remains consistent (which is not always the case)was that siting issues affected readings so that thermometers read to high. Neither Watts nor anyone among his cheerleading crowd ever attempted to do a real data analysis to verify the hypothesis. One of his readers, however, tackled the problem as sson as enough stations were sampled (John V). He evidently found out that the hypothesis was not verified by data analysis and endured so much malice at Watts's site that he didn't post there any more. Further analysis was done by NOAA once enough stations were sampled so that no regional bias could possibly affect the results, and the results were exactly the same. The very premise for the existence of Watts' blog has been invalidated numerous times.
    0 0
  11. Philippe, thanks for the clarification - I should have made it clear my comment was directed at some of the comments, not at the post itself. Everyone knows that the UHI exists, the IPCC has it at 0.006'C per decade. Perhaps that's correct, but the Fujibe paper and the Ren paper do question that number. Anyway, I'm off topic, as this post is about the effect of microsite issues on measurement.
    0 0
  12. Markus (Post #1) is already aware that he would be ill-advised to hold his breath waiting for an acknowledgment from Anthony Watts that the U.S. surface temperature data set apparently does not show a systematic bias toward warming. In fact, in the overall scheme of things, I’d wager it’s more likely that Watts would attempt to stop publication of the paper than to make such an admission. This issue has much broader implications, however, that extend well beyond the reliability of this particular data set, and bears on the entire debate over anthropogenic climate change (ACC). Skeptics such as Watts, Joseph D’Aleo and others are well within their rights to question the validity of the data, particularly considering the poor condition and non-ideal location of many of the measurement stations. More than that, it’s their duty (duty…duty…duty!) to raise skeptical criticisms, just as it is the duty of climate scientists to address reasonable concerns. The advancement of science depends on it. The success of this approach demands, however, that skeptical hypotheses be: a) testable, and b) potentially refutable. If not, then they fall into the domain of ideology, not science, and can never be considered anything more than unsubstantiated conjecture. Skeptics feel they’ve done their job merely by raising questions and doubts, while forgetting the essential next step of hypothesis testing. Sadly, past experience shows that in the current "debate", most arguments against anthropogenic climate change are effectively irrefutable, no matter how much evidence is brought to bear. Worse, the premise that ACC is wrong provides the touchstone by which all evidence is measured. Evidence that appears to support ACC is inferred to be wrong; evidence that appears to refute ACC is inferred to be valid. At the same time, the new re-assessment of the data by Menne et al. gives all of us a greater level of confidence in its reliability. (Readers may also be interested to read this analysis of the NOAA & NASA data: http://www.yaleclimatemediaforum.org/2010/01/kusi-noaa-nasa/ In fairness, we have to acknowledge the important role that skeptics have played in this process. It’s a shame if Watts and others are unable to derive any satisfaction from their efforts, even if the rest of us can. Most likely they’ll just keep plugging away, trying to prove what they already ardently believe. The surface temperature data set does, indeed, pose a “challenge” to put it mildly. While the temptation is there to just chuck the entire lot, we can’t afford to do that, as the results are too important. The only other option is to try to make the best use of the available information, by removing as much of the error as possible without introducing bias. This may entail eliminating some stations, and making some adjustments to some data, where warranted. This is what the researchers at the National Climate Data Center have been trying to do in good faith. They deserve our appreciation, and that of Mr. Watts as well!
    0 0
  13. CoalGeologist, people who raise doubts about the science, without seeking to prove that those doubts are valid, are not true skeptics-they're denialists-an important distinction. Lindzen, for example, is a skeptic-he doubted that global warming would be as extreme as predicted (due to the Iris Effect) & sought to prove it-so he's a skeptic. On the matter of bias in US temperature records, Watts has behaved as a true skeptic-up to a point-but if he now refuses to publish the results of this study, then he proves himself to be just another denialist. His role as a denialist is already proven, however, in the way he behaves on other matters related to Anthropogenic Global Warming.
    0 0
  14. stevecarsonr, my apologies-I misspoke before. I implied that the Urban Heat Island was a myth-that wasn't my intent. What I meant was that Urban Heat Islands being the primary cause of global warming (rather than CO2 emissions) was an Urban Legend. This paper seems to give added weight to the Urban Legend status of the view that poorly sited measuring stations, alone, are capable of producing a +0.16 degree change per decade in average global temperatures!
    0 0
  15. HumanityRules-also my apologies. I wasn't actually suggesting that the denialists *invented* the urban heat island (even if that's how it comes across), but they have exploited it *mercilessly* to try & undermine the credibility of the surface temperature record-even long after study after study had shown that (a) the bias wasn't as strong as suggested (b) the bias often gave cooler results than for nearby rural areas (c) researchers always adjusted for the bias & (d) that the surface record was closely correlated to the record from satellites. I doubt that this latest paper will silence their misuse of UHI for ideological purposes-even when its based on the work of one of their own!
    0 0
  16. Marcus, it's not my blog, added to which I'm new here - so apologies if my comment is out of order - but I don't understand the attraction of the labeling of others. (And I've seen it a lot at other blogs). You might characterize their opinion, that's certainly helpful. But "skeptic"/"denialist" seems like a classification of character - or maybe assassination of character. Now you might say the 2 people you are talking about are a distinguished physics professor and a meterologist, so take the following as a general comment that doesn't apply to them as to the specifics, but maybe the general concept does.. Some people don't understand the radiative transfer equation (in fact, I've just realised that maybe I don't understand it properly, maybe someone can help with my totally off-topic question) because they don't have a physics background. So they don't understand how CO2 can impact temperature. Does this make them "a denialist"? Or someone who doesn't understand radiative physics? People are free to call them whatever they like, but one comment I would make it that the more personal attacks thrown the less likely people with questions are to sit down and try and understand a complex subject. And it is complex. And the scientific method is not natural and instinctive.
    0 0
  17. Gentlemen You really ought to read the methods used before you gloat. The individual station anomaly measurements were based on each stations "1971-2000 station mean". See where the document states: "Specifically, the unadjusted and adjusted monthly station values were converted to anomalies relative to the 1971–2000 station mean." In other words, the only thing this study measures is the difference in instrument error at each station. The absolute error occurring at individual stations because the station had not been properly located is not measured. A poor station with an absolute temperature error of +5 degrees C still has a bias error of +5 degree C - no matter what the variation occurring due to instrumentation type. I'm a chemical engineer with U.S. government and 20 years of research experience in various areas including environmental mitigation. If one of my phD's came to me with this nonsense, I'd fire him on the spot. Sorry boys, you are going to have to better than this. Dave
    0 0
    Response: Whenever you look at a graph of global temperature, invariably you're looking at "temperature anomaly" (the change in temperature), not absolute temperature. As NASA puts it, "the reason to work with anomalies, rather than absolute temperature is that absolute temperature varies markedly in short distances, while monthly or annual temperature anomalies are representative of a much larger region". It's the change in temperature (eg - the trend) that is of interest and the analysis in Menne 2010 determines if there is any bias in the trend due to poor siting of weather stations.
  18. Steve Carson. I'm no expert in climatology, but if I read a paper which claimed that surface temperatures had been falling (not rising) for the last 30 years, then I'd seek independent verification from other sources before I accepted or dismissed the claim-that's what makes me a Skeptic (& a scientist). A denialist, by contrast, will automatically dismiss any evidence that doesn't fit their ideology-without independent verification-no matter how strong the evidence is (yet they still demand ever more evidence-even though they'll dismiss that too). If it helps, the other side contains what I call the "True Believers"-they accept the theory of global warming because someone they admire &/or want to believe tells them so-without independent verification. Personally, I have no time for denialists or true believers, but instead seek independent verification of every claim & counter-claim being made. Its always important to think for yourself rather than blindly accept the claims of people who might have a vested interest. Hope that makes more sense.
    0 0
  19. Swerving off topic so possibly may never see the light of day, but further to remarks on skepticism versus denial, etc., English is a rich language and there's no need to use a single word to describe a plethora of approaches. Doubter, contrarian, skeptic, denier, they're all different in meaning and need to be applied individually. "Faithful" would be a better word for some, for that matter, seemingly detached from the material world. My limited experience w/participating in discussions on this topic tells me I'm generally far too hasty in categorizing, to the point where I've already had to resort to apology too often, enough to make me more cautious about committing accidental slurs. As is said, discretion is the better part of valor.
    0 0
  20. Kforestcat, I think you may be confusing two things here - bias error and probabalistic noise. The paper makes it clear that the unadjusted curves represent the measurements before known bias errors are removed, while the "adjusted" curves are after the bias errors have been corrected. Conversion to an anomaly is effectively the same as normalization, and the purpose is the same. Both serve to accentuate the part of the data that you care about. I do both regularly in my professional field of electrical engineering, especially when I'm interested in understanding the nature of noise plaguing my circuitry. Finally, what Watts et al are essentially saying is that heat islands, in this case caused by electrical transformers, waste treatment plants, air conditioners, or pavement, have made the global temperature record unusable. This paper points out that Watts is incorrect, but it's not the first paper to do so by any means. The following paper showed that well established urban areas had the exact same trends as rural areas, but with a removable warm temperature bias: http://www.agu.org/pubs/crossref/2008/2008JD009916.shtml To use an analogy, if a trampoline can get you 10 feet into the air out on a farm, there's every reason to believe that it'll get you just 10 feet into the air if you move it into a city.
    0 0
  21. Gentlemen I'm fully aware of how anomaly data is used ( having used it in my own research) and I know full well what can go awry in the field experiments. We are talking about every day instrument calibration and QA/QC - this is not rocket science. I firmly maintain the Menne 2010 paper is fundamentally flawed and entirely useless. NASA's individual station temperature readings are taken in absolute temperature (not as an anomaly as you have suggested). The temperature data is reduced to anomaly after the absolute temperature readings for a site are obtained. For example see, the station data Orland (39.8 N, 122.2 W) obtained directly from the NASA's GISS web site. The temperatues are recorded in Annual Mean Temperature in degrees C - not as an anomaly as you have suggested. (Tried to attach a NASA GIF as visual aid -but did not succeed). Bottom line. Menne has to have (and use) absolute temperature data to get the 1971-2000 mean temperature and then divide the current temp with the mean to get the anomaly. We are back to the same problem - Menne is measuring instrument error - he is not measuring error resulting from improper instrument location. The Menne paper is absolutely useless for the stated purpose. Anyone who actually collects field data, I have, knows they are going to immediately run into two fundamental problems when an instrument is improperly located. 1) they are not reading ambient air temperature and 2) neither temperature readings nor the anomaly can be corrected back to a true ambient because other factors are influencing the readings. For example: Suppose we have placed our instrument in a parking lot. Say the mean 1971-2000 temperature well away from the parking lot is 85F; but the instrument is improperly reading a mean of 90F. Now on a given day, say the ambient temp is 93 but your instrument is reading 105F (picked up some radiant heat from a car). Ok our: Actual anomaly is 93F - 85F = 8F; Instrument anomaly is 105F - 90F = 15F. The data is trash. There is simply no way to recover either the actual ambient temperatures nor an accurate anomaly reading. What you are missing is that an improperly placed instrument is reading air temperatures & anomalies influenced by unnatural events. The readings bear no relationship to either the actual temperature nor the actual anomaly - the data's no good, can't be corrected, and will not be used by a reputable researcher. Finally, it's not entirely surprising that Menne finds a downward bias in his individual anomaly readings at poorly situated sites. Because: 1) a poorly located instrument produces a higher mean temperature; hence, the anomaly will appear lower; and 2) generally there's a limit to how hot an improperly placed instrument will get (i.e. mixing of unnaturally heated air with ambient air will tend to cool the instrument - so the apparent temperature rise is lower than one might expect). Had Mennen (NASA) actually measured both absolute temperature and calculated anomaly data using instrumentation at properly setup sites, within say a couple of hundred feet of the poor sites, as a proper standard to measure the bias against - our conversation would be different. As it stands Menne's data is useless nonsense and not really worth serious discussion. Dave
    0 0
  22. Kforestcat, of course the temperature stations produce absolute temperatures as their "raw" data rather than as anomalies from a baseline. I have never seen anyone claim otherwise. You are misreading quite drastically. The baseline against which the anomalies are computed, is the average temperature for that specific locality across whatever time range has been chosen as the baseline. Each station has its own, local, baseline computed. Then each individual temperature reading from that one given station is differenced from that baseline for that one given station. The result is a difference of that one reading, from that tailored baseline. That procedure is done separately for each individual temperature reading, each against its own individual, tailored, baseline. It is a simple mathematical transformation that has nothing to do with instrument error and nothing to do with instrument calibration. It is a simple re-expression of each individual temperature reading that preserves all changes from the baseline temperature. The resulting collection of individually transformed temperatures is the collection of "raw" anomalies. Those are the "raw" data that you see being discussed.
    0 0
  23. Kforestcat, in your example you wrote "Say the mean 1971-2000 temperature well away from the parking lot...." But that's not of interest. Instead, the temperature on that given day, from that parking-lot-situated instrument, is differenced from the average temperature across 1971-2000 of that same instrument.
    0 0
  24. Regardless of site, we see an obvious rising temperature gradient from 1980 to 2009 based on a line of best fit. However, ‘eyeballing’ the unadjusted data suggests a striking fall based on a line of best-fit beginning with an anomalously high 1998 to 2009. Now we certainly don't want to cherry pick. The 1998 data was attributable to a very large El Nino. However, following on from the preceding post (‘The chaos of confusing the concepts’) with its discussion of the Lorenz attractor, I find myself wondering whether we may indeed be seeing evidence of greater inherent unpredictability than we commonly suppose. Eleven years after all seems a long period, especially when we consider the preceding data set covers eighteen years. Should we be considering the two periods as one segment? Alternatively, should we be considering these periods as two distinct segments and asking why 1998 produced such a high El Nino (followed by a relatively warm period) and why 2007 – 2009 are producing a much lower gradient? Moreover, is this gradient likely to continue? I think the question of site location is clearly a furphy given the broad consistency between better and not so well located sites. However, deciding which periods we select to measure trends is of much more fundamental importance given the arbitrary nature of lines of best fit. Otherwise, we risk failing to ask obvious questions.
    0 0
  25. Kforestcat, I'm sorry, but you're off in the weeds on this one. What you describe with your pavement example is an example of signal + bias + noise. Because the instrument's location is constant, we can eventually come up with a correction mechanism to remove the bias from the data. That leaves us with signal + noise. Removing the noise is simple filtering, of which averaging is one variety. Mathematically, averaging a signal removes noise (increases the signal-to-noise ratio) at the rate of the square root of the number of samples. Averaging daily samples over the course of a week increases the SNR by nearly 3 over any single sample. So if we picked up thermal noise from a car one day, then we merely have to average that data point with others from the same instrument in order to dramatically reduce the impact of that noisy sample on the overall data. I'll grant you that, if you only have a single data point, biases and noise on that data point will be a major problem. But that's not the case with the temperature record.
    0 0
  26. kforestcat writes: "I'm fully aware of how anomaly data is used ( having used it in my own research)" I'm not sure you actually do understand this, because your comments still show the same kinds of errors and confusion. "NASA's individual station temperature readings are taken in absolute temperature (not as an anomaly as you have suggested)." NASA doesn't take temperature station readings, and nobody has suggested that the temperature sensors measure anomalies directly. "Menne has to have (and use) absolute temperature data to get the 1971-2000 mean temperature and then divide the current temp with the mean to get the anomaly. We are back to the same problem - Menne is measuring instrument error - he is not measuring error resulting from improper instrument location." That is very confused. The temperature anomaly is the current daily (or monthly) temperature minus the mean temperature on the same day (or month) during a given reference period. You don't "divide" any temperatures. And Menne et al. are not measuring "instrument error". They are analyzing measurements of temperature as a function of site quality in order to determine the difference in temperature trends between well-sited and poorly-sited stations. "Actual anomaly is 93F - 85F = 8F; Instrument anomaly is 105F - 90F = 15F. The data is trash. There is simply no way to recover either the actual ambient temperatures nor an accurate anomaly reading. What you are missing is that an improperly placed instrument is reading air temperatures & anomalies influenced by unnatural events." You still completely fail to understand what's going on here. Menne et al. are taking the temperature data and grouping them into categories based on the site quality. They then determine the difference in long-term trends between well-sited and poorly-sited stations. In the raw, unadjusted data, poorly-sited stations tend to have a slightly lower trend than well-sited stations. The network homogenization and adjustment process brings poorly-sited stations into closer agreement with well-sited stations. "The readings bear no relationship to either the actual temperature nor the actual anomaly - the data's no good, can't be corrected, and will not be used by a reputable researcher." That is just bluster. What the analysis shows quite clearly is that if anything, poorly-sited stations on average underestimate the warming trend, but that the network adjustment process is able to successfully compensate for this effect. And even if you were reluctant to accept that, the close agreement between in-situ surface temperature and satellite microwave temperature retrievals from the lower troposphere suggests that the surface temperature record is realistic. "Finally, it's not entirely surprising that Menne finds a downward bias in his individual anomaly readings at poorly situated sites. Because: 1) a poorly located instrument produces a higher mean temperature; hence, the anomaly will appear lower; " Huh? Again, this makes no sense. If a sensor always reads 5C too high, its anomaly will be exactly the same as if it were perfectly sited. If a sensor's environment changes such that the current temperature is biased high relative to the period of record, then it will have a positive anomaly, not a negative one. "and 2) generally there's a limit to how hot an improperly placed instrument will get (i.e. mixing of unnaturally heated air with ambient air will tend to cool the instrument - so the apparent temperature rise is lower than one might expect)." That is both confused and irrelevant to the paper at hand. "Had Mennen (NASA) actually measured both absolute temperature and calculated anomaly data using instrumentation at properly setup sites, within say a couple of hundred feet of the poor sites, as a proper standard to measure the bias against - our conversation would be different." (1) Menne et al. work for NOAA, not NASA, and the paper being discussed here is about NOAA's temperature data. (2) You still seem confused about the relationship between measured temperature data and calculated temperature anomaly. (3) The entire point of this paper is to compare poorly-sited and well-sited stations. (4) By doing this comparison using trends in the anomaly rather than using the absolute temperatures, there's no need to compare stations within "a couple of hundred of feet" of each other. "As it stands Menne's data is useless nonsense and not really worth serious discussion." Again, that is just bluster. It sounds to me like you don't understand the subject but are deeply invested in casting doubt on it.
    0 0
  27. " A poor station with an absolute temperature error of +5 degrees C still has a bias error of +5 degree C - no matter what the variation occurring due to instrumentation type." We're interested in trends, so a constant bias has no effect, nor does the choice of baseline from which to compute the anomaly. For any bias B, and any two temperature reading at points in time N0 and N1, (N0-B) - (N1-B) = N0 - N1. And you can extend that into any statistical trend analysis taken over a time series N0 ... Nn. "I'm a chemical engineer with U.S. government and 20 years of research experience in various areas including environmental mitigation. If one of my phD's came to me with this nonsense, I'd fire him on the spot. " I could make a snarky statement about 9th grade algebra students but I'll withstrain myself.
    0 0
  28. Or maybe even restrain myself :)
    0 0
  29. Interesting debate. So we are talking measuring trends vs actual data, yes? But this does not really answer the Watts paper http://wattsupwiththat.files.wordpress.com/2009/05/surfacestationsreport_spring09.pdf To me Kforestcat, along with Watts paper, makes sense. I cant see the point of the Menne exercise - why bother with a trend when you could measure how good the station was at measuring temperature. Why not put the army of volunteers to good use - how long would it take?
    0 0
  30. jpark at 06:37 AM on 24 January, 2010 It's not actually a debate, instead repeated attempts at explanation. Try this one, at home if you like but it's so simple you'll probably find words do the job: Set up a thermometer in a large room where the temperature is steady. Let the thermometer stabilize at ambient temperature. Now, turn on a small lamp next to the thermometer, close enough to warm it a bit. You'll see an immediate bias in the reading given by the thermometer; the reading will be higher than ambient temperature in the room. Let the thermometer stabilize again. Now raise slowly raise the temperature of the room. The thermometer will still register the increase in the temperature of the room. We've learned that bias does not make it impossible to extract a trend in temperature. It's really -that- simple. Not so hard, really, but easy to lose in a detailed technical explanation. To me it seems what we have here after all the hat and light is stripped away is the famous "failure to communicate".
    0 0
  31. Hi Doug! Many thanks, nice explanation. - I do understand the paper but still feel it does not, like a lot of posts here, answer the quite basic Watts question of how accurate the stations are. http://wattsupwiththat.com/2010/01/23/quote-if-the-week-27/#more-15561 And the picture tells a 1000 words - how do you convince people of global or even just US warming when you get to see a weather station next to a/c. It has been a bad week for AGW.
    0 0
  32. I'm not a scientist, but this topic fascinates me. I do have 30 years of professional experience to help guide me. When a colleague speaks in abruptly dismissive terms, claiming something is "useless," "trash," or "not really worth serious discussion" I pay attention, but my guard goes up. My years of experience have taught me to listen, but be skeptical. I have rarely found that such a tone is warranted. Here again, I appreciate the careful explanations by people who have responded. I am not a blind believer, but I do have confidence that serious professionals are sincere and careful in their effort, and are correct more often than not. I think that the argument that temperatures are rising is well backed by the loss of sea ice extent, and especially the rapid loss of multi-year ice in the past couple years. The next few years may be telling.
    0 0
  33. jpark, I doubt that Watts et al photographed all of the stations, including well sited ones. This study shows that both well sited and poorly sited stations show the same basic trend, and that the bias in poorly sited stations is to cooler temperatures. What could be clearer than that? You haven't seen a bad week in AGW yet. You may well see many in your lifetime. I hope we don't.
    0 0
  34. Gordon, many thanks. I think what you say is wise, But do look at the Watts report - lots of stations, lots of pictures ( worked for Al, it might just work for Ant) http://wattsupwiththat.files.wordpress.com/2009/05/surfacestationsreport_spring09.pdf
    0 0
  35. As others have said the bias does not affect the trend. "A rising tide lifts all boats." Increased GHGs are the rising tide. Watts refuses to admit the obvious. In fact, even today he has the following post: http://wattsupwiththat.com/2010/01/23/sanity-check-2008-2009-were-the-coolest-years-since-1998-in-the-usa/ He is using US data to try to cast doubt on global data. He knows better but loves the attention from his misguided followers. I used to post there as the loyal opposition but it ended up being a huge waste of time.
    0 0
  36. Thanks Prof! Again very helpful but to me a trend is not at issue. I am a newbie here and like many this area of science has only become a really hot (pardon the pun) topic for me since climategate. Before then I accepted the general consensus. Most in the blogosphere seem to believe in global warming - it is the extent and the 'unprecedented' nature of the warming that I think (from what I have read on blogs so far) is the issue, which is of course linked to the anthropogenic part. This means actual temps do matter and trends in this particular instance dont, to me at any rate. If it is getting a bit hotter then, well that is not too bad, climate does tend to do that. But if it is getting amazingly hotter then, of course, we are all going to be in big trouble. So are the temps showing something dangerous or something not so dangerous I read this "Why Hasn't Earth Warmed as Much as Expected? New Report on Climate Change Explores the Reasons" from Science Daily. I think you can understand my layman's puzzlement. http://www.sciencedaily.com/releases/2010/01/100119112050.htm Apologies if that is off topic.
    0 0
  37. Obsessively repeating the same concept is the standard tool people in the media use to make it true. Anthony Watts is pretty good at it and the fact that even if true it does not have any pratical impact has no importance for him and his fellows. They'll stubbornly keep repeating "Poor siting! Poor siting!" ad infinitum.
    0 0
  38. Reading some of the posts here is incredibly frustrating because it clearly demonstrates the stunning success Watts et al. have had in confusing and brainwashing people (even well educated professionals it seems)to the point where it is impossible to explain a simple concept of a temperature anomaly to them. I was going to chime in and try to dispel some of the confusion, but others have repeatedly and clearly explained the facts only for those facts to repeatedly fall upon deaf ears. What I will add, is that the Menne et al. study needed to be done and their results are incredibly important. Their results also represent the final nail in the coffin for the complaints from Watts et al. as to the validity of the US SAT record. There is simply no dobting the validity of the SAT record anymore, but I doubt this study will discourage the contrarians and denialists from perpetuating and rehashing old myths. Prof Mandia re #35, I too once tried to explain the science with the folks at WUWT, and it was a waste of time. Watts knows his audience and plays to that; he is very good at telling them what they want to hear. He is also guilty of confirmation bias and ignoring the inconvenient facts regarding AGW. Anyhow, I do hope that some of the misguided posters here represent the views of people who are in the minority, b/c if they represent a much larger segment of the populous then we have a serious problem on our hands in terms of communicating the science. Why is it so much easier to disseminate misinformation than the basic facts? Maybe someone with some time can show some schematics illustrating how one obtains anomaly values from a temperature record, and why systematic bias does not affect the trend? A picture is oftentimes far more convincing and informative than even the most carefully chosen words. PS: Actually those in denial are having a bad decade-- 2009 second warmest year on record globally, first decade of naughts warmest on record globally, warmest year on record in S. Hemisphere (lots of heat stored in the vaste southern oceans), continuing acceleration of rate of loss of summer Arctic sea ice and glaciers, PIG glacier in WAIS found to have exceeded its tipping point, and for what it is worth, January 2010 warmest lower trop. temps in the satellite record despite extremely cold weather in Eurasia and portions of N. America. The list goes on.....
    0 0
  39. Jpark "This means actual temps do matter and trends in this particular instance dont," The actual temperatures form part of a long-term trend. You can't have a trend in a time series without either increasing or decreasing time series of temperatures. Moreover, those tmepratures do not have to increase montonically to get a positive trend as illustrated by the surface air tmeprasture records. The long term temperature trend (globally) is about 1.7C warming per century, and yes, that is actually something to worry about. Regarding "why we have not warmed as much as we should have". You are probably referring to the work of Scwartz that is aboutt o be publishe din J. Climate. Perhaps John can again (Schwartz has done this before) refute the work of Schwartz et al. Jpark, be wary of site slike WUWT, their goal is to confuse. Really it is just that simple, and it is cleverly done under the guise of "science" and the pursuit of "truth". That is what makes the misinformation there seem so compelling. The long term observed warming trends is consisent with the projections made by the IPCC. Look here: http://tamino.wordpress.com/2010/01/13/models-2/ and here http://tamino.wordpress.com/2009/12/07/riddle-me-this/ and here http://tamino.wordpress.com/2008/01/31/you-bet/ I really encourage you to actually read the above articles carefully.
    0 0
  40. Albatross - I think an illustration on temp anomaly might be a good idea. Ok I will give this one more go because I think you guys might be able to give me an answer and you haven't yet (apart from Kforestkat) Here is the problem: the world is getting hotter - I think we all agree - but last year we found out that CRU scientists made a hash of doing the data. However kindly you read the leaked emails you realise this was not good science. Then Copenhagen fails. Then this week Pachauri gets it in the neck for getting the Himalayan glacier date wrong and putting pure speculation in the IPCC report (apparently it was not the only error) and an error that had significant financial consequences. So when Watts puts out a report showing images of severely compromised temp stations and Menne replies with 'trends' people like me say...'er so what? What does a trend mean, I want to know whether the temp stations work or they are being lovingly heated by a/c units". Because if those temps/trends are slightly higher than they should be and so, in reality, only slightly higher than older temp station data, or even older historic data then, yes, we have global warming but not very much - which is what the report at Science Daily says. But of course I may be missing something...
    0 0
  41. Albatross - thanks for the links - I will read. Carefully.
    0 0
  42. jpark, "if those temps/trends are slightly higher than they should be and so, in reality, only slightly higher than older temp station data, or even older historic data then, yes, we have global warming but not very much" Exactly, but it ain't so. In principle it might be a resonable concern, in practice it does not stand up an in depth analisys. Remember, people working on it check the readings for possible biases/errors; something may slip through the check but, well, just some. And unless you belive in the bad intentions of the researchers, errors and biases (plural) tend to average out. Don't be confused by absolute temperature and anomaly. The former is more intuitive given that it's what we feel. The latter has the advantage of being more stable and correlated over long distance and time, then more easily shows an underlying trend, which is what we are interested in.
    0 0
  43. jpark at 07:05 AM on 24 January, 2010 Hi Doug! Many thanks, nice explanation. - But still not good enough, and I'm sorry. "I do understand the paper but still feel it does not, like a lot of posts here, answer the quite basic Watts question of how accurate the stations are." When you use the phrase "how accurate the stations are" I think it betrays that you don't understand the paper. It does not matter at all if the stations are accurate. Their utility for telling accurate absolute temperature from day to day is entirely separate from their utility for revealing a climatic trend. All that matters for extracting a trend is whether or not there's a unidentified longitudinal change of bias in measurements resembling a trend in temperature. More, that unidentified longitudinal change must be approximately the same for a multitude of stations. As it happens, there is no unidentified longitudinal bias change that meets that requirement, but there is a -know- reason for observing a trend, namely a change in climate.
    0 0
  44. jpark: You're starting to get all mixed up here. "...last year we found out that CRU scientists made a hash of doing the data." Wrong. Be careful, there are a number of people with an unhealthy obsession with old email, to the point they've actually set up a web site all about it and apparently spend their days sifting through this stuff. There's really no "there" there. "Then Copenhagen fails." And that does not have anything to do with surface stations in the US. "Then this week Pachauri gets it in the neck ..." Again, nothing to do with this topic. "What does a trend mean..." A trend tells you useful things, such as whether you can expect your coffee to ever brew. "Because if those temps/trends are slightly higher than they should be and so, in reality, only slightly higher than older temp station data, or even older historic data then, yes, we have global warming but not very much..." There you go! That's the useful part! The trend provides confirmation of theory via observation, validation of models, etc. And that's why Watts et al are so determined to distract you from the importance of trends. Easy once you go through the steps, plus remember these folks are doing the same thing you used to see in word problems: throwing a lot of chaff in the air to confuse you so you can't come to a useful conclusion. Don't let yourself flunk the test.
    0 0
  45. I've been following these threads and seeing how everything is going back and forth. Considering when somebody turns on the news and watch the weather report, we are presented with it in absolutes. So to an average laymen I could see how it seems to make more sense to want to see the data presented like that. And considering that the pdf on WUWT is very pretty and professional looking I could see how people will believe it. (He must have an army of dedicated followers, no wonder he doesn't want to report anything different, he will lose his crown) It wasn't until I started playing around with temperature data myself (DIY-Statistics) that I could understand it a whole lot better!! Is there any room for a post John or Mark on "anomalies verses absolute" or "this is what a raw reading looks like, this is what has to be done to extract sense from it..." (with lots of pictures of course, people like their pictures)?
    0 0
  46. jpark. Trends are very important because they tell us a lot about the *rate* of change. This is especially important if we can compare it to rates of change in the past. You see, though climate has changed in the past, all the available evidence suggests that it has *never* changed as rapidly as it has in the last 30-60 years, in spite of a relative lull in Total Solar Irradiance. That is why climatologists are so concerned, in spite of the efforts of people like Watts to confuse the issue.
    0 0
  47. Deniers are very good at confusing lay people in the use of absolute temperature and anomalies. For example Joseph D’Aleo and E. Michael Smith have accused "NOAA researchers of strategically deleting cherry-picked, cooler-reporting weather observation stations from the temperature data". They say that by ignoring these cooler stations the global temperature is artificially raised over what it would be if the stations were included. However, if you actually look at where the deleted or ignored stations are they are in areas of the world which are experiencing much faster rates of temperature increase than average (northern Canada, northern Russia). Thus their omission is actually lowering the global average, the exact opposite of what D'Aleo and Smith are saying.
    0 0
  48. Albatross - guys, thanks. The trend and climate models are great. I will study them some more, and of course I find them persuasive but my immediate reaction is 'oh, more graphs with dots on, is that it?' Others here have parried my question rather than answered it. Is the actual data duff or not? If the data going into all the trendy models is bad then we, the public, will simply dismiss the model. The Met Office in the UK got their predictions (and models) badly wrong this year while, to their utter horror, a chap called Piers Corbyn got it right. So much so that the BBC are considering using a different company. I know this has to do with 'weather' rather than 'climate', but let's face it if you cant predict one then the other looks fanciful. The UK Gov are setting up a parliamentary inquiry into the CRU leaked emails - this is good because it should be thorough and open. But it does mean, Doug, that you cannot say there is nothing to the emails and everything is fine. The IPCC report is, I am afraid, also important. If it includes rather wild speculation about Himalayan glaciers then the whole report looks rather suspect. Ian says that deleted stations are showing greater increases of warming - then why the heck are they not included in the data? Why do we have this adjusted/deleted/averaged/smoothed picture of what is happening - why not real/complete/comprehensive? So back to the topic here - if Menne's paper just tells us that the surface station data does nice trends then I, for one, am still left scratching my head. The case for catastrophic global warming seems too dependent on 'managed data' to me. After the Y2K bug, the sub prime mortgage/financial crisis there is good reason to be sceptical. This is very very poor PR - there has to be better than this.
    0 0
  49. jpark, can you be a little more specific about how the Met Office "got it wrong"? If you're talking about the recent cold snap, that was caused by an unforeseen change in the Arctic Oscillation, & has absolutely nothing to do with broader warming trends. What Menne's paper shows is that, in spite of the claims by people like Watts, so-called "poorly sited" weather stations are showing only a negligible difference in both their minimum & maximum temperature readings-over time. This therefore means that poor siting of US weather stations cannot be used to explain the global warming trend of the last 30+ years. This is hardly news, as satellites have shown an almost identical warming trend over the 30 years they've been taking readings. Your claim about "managed data" is meaningless, as *all science* is dependent on the manipulated-or management-of raw data. To try & equate "management" or "manipulation" with fraud is to essentially impugn the entire scientific establishment. Also, as someone who actually *knows* people in the IT industry, I can assure you that Y2K was not a hoax (though some news agencies deliberately overstated the threat). Had nothing been done about it, many industrialized nations would have been disrupted for days-if not weeks. Last of all, many of the people who helped create the sub-prime mortgage crisis walk in the same circles as those who're pushing the denialist cause in the media. Their motives are also identical-*profit*. That's why I'm so skeptical of the denialist case.
    0 0
  50. More points jpark. Inquiry or not, the denialists who hacked & distributed the CRU e-mails have now had close to 3 months to find something truly damning about the way in which CRU has collected &/or manipulated the data. That they've not presented any case of fraud in this time strongly suggests that this is because no evidence exists to prove it. I'd personally like to know when there is going to be an inquiry into who hacked CRU-& who paid them-& when those responsible will be brought to justice (last time I checked, Hacking was a MAJOR CRIME in most constituencies). That such an inquiry hasn't taken place suggests that some very powerful vested interests were behind the hack. One wonders what you want, jpark? If graphs showing the warming trend don't suit you, then what about the images of the Earth covered in ever greater shades of orange & red showing the extent of warming over the last 30 years? Graphs remain the very best way of showing how the minimum & maximum temperature anomalies for each decade have changed. As to why certain stations are omitted, it might be for any number of reasons. Maybe local conditions meant the station was off-line for too many days out of a year, or maybe some localized event caused the station to become an obvious outlier. Maybe they simply had enough replicate data points, from a specific region, to get an average with a sufficiently small margin of error. The point is that the deletion of a handful of stations across the globe isn't suddenly evidence of a conspiracy. What is evidence of a conspiracy, though, is how many of the official denialist groups & individuals have strong ties to the mining & fossil fuel industries.
    0 0

1  2  3  4  Next

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



The Consensus Project Website

THE ESCALATOR

(free to republish)


© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us