the Air Vent

Because the world needs another opinion

Considered Critique of Berkeley Temp Series

Posted by Jeff Id on November 3, 2011

I will leave this alone for another week or two while I wait for a reply to my emails to the BEST group, but there are three primary problems with the Berkley temperature trends which must be addressed if the result is to be taken seriously.  Now by seriously, I don’t mean by the IPCC which takes all alarmist information seriously, but by the thinking person.

1 – Chopping of data is excessive.   They detect steps in the data, chop the series at the steps and reassemble them.   These steps wouldn’t  be so problematic if we weren’t worrying about detecting hundredths of a degree of temperature change per year. Considering that a balanced elimination of up and down steps in any algorithm I know of would always detect more steps in the opposite direction of trend, it seems impossible that they haven’t added an additional amount of trend to the result through these methods. Steve McIntyre discusses this here. At the very least, an examination of the bias this process could have on the result is required.

2 – UHI effect.  The Berkeley study not only failed to determine the magnitude of UHI, a known effect on city temperatures that even kids can detect, it failed to detect UHI at all.  Instead of treating their own methods with skepticism, they simply claimed that UHI was not detectable using MODIS and therefore not a relevent effect.

This is not statistically consistent with prior estimates, but it does verify that the effect is very small, and almost insignificant on the scale of the observed warming (1.9 ± 0.1 °C/100yr since 1950 in the land average from figure 5A).

This is in direct opposition to Anthony Watts surfacestation project which through greater detail was very much able to detect the ‘insignificant’ effect.

Summary and Discussion
The classification of 82.5% of USHCNv2 stations based on CRN criteria provides a unique opportunity for investigating the impacts of different types of station exposure on temperature trends, allowing us to extend the work initiated in Watts [2009] and Menne et al. [2010].
The comparison of time series of annual temperature records from good and poor exposure sites shows that differences do exist between temperatures and trends calculated from USHCNv2 stations with different exposure characteristics. 550 Unlike Menne et al. [2010], who grouped all USHCNv2 stations into two classes and found that “the unadjusted CONUS minimum temperature trend from good and poor exposure sites … show only slight differences in the unadjusted data”, we found the raw (unadjusted) minimum temperature trend to be significantly larger when estimated from the sites with the poorest exposure sites relative to the sites with the best exposure. These trend differences were present over both the recent NARR overlap period (1979-2008) and the period of record (1895-2009). We find that the partial cancellation Menne et al. [2010] reported between the effects of time of observation bias adjustment and other adjustments on minimum temperature trends is present in CRN 3 and CRN 4 stations but not CRN 5 stations. Conversely, and in agreement with Menne et al. [2010], maximum temperature trends were lower with poor exposure sites than with good exposure sites, and the differences in
trends compared to CRN 1&2 stations were statistically significant for all groups of poorly sited stations except for the CRN 5 stations alone. The magnitudes of the significant trend differences exceeded 0.1°C/decade for the period 1979-2008 and, for minimum temperatures, 0.7°C per century for the period 1895-2009.

The non-detection of UHI by Berkley is NOT a sign of a good quality result considering the amazing detail that went into Surfacestations by so many people. A skeptical scientist would be naturally concerned by this and it leaves a bad taste in my mouth to say the least that the authors aren’t more concerned with the Berkley methods. Either surfacestations very detailed, very public results are flat wrong or Berkeley’s black box literal “characterization from space” results are.  Someone needs to show me the middle ground here because I can’t find it.

I sent this in an email to Dr. Curry:

Non-detection of UHI is a sign of problems in method. If I had the time, I would compare the urban/rural BEST sorting with the completed surfacestations project. My guess is that the comparison of methods would result in a non-significant relationship.

3 – Confidence intervals.

The confidence intervals were calculated in this method by eliminating a portion of the temperature stations and looking at the noise that the elimination created. Lubos Motl described the method accurately as intentionally ‘damaging’ the dataset.  It is a clever method to identify the sensitivity of the method and result to noise.  The problem is that the amount of damage assumed is equal to the percentage of temperature stations which were eliminated. Unfortunately the high variance stations are de-weighted by intent in the processes such that the elimination of 1/8 of the stations is absolutely no guarantee of damaging 1/8 of the noise. The ratio of eliminated noise to change in final result is assumed to be 1/8 and despite some vague discussion of monte-carlo verifications, no discussion of this non-linearity was even attempted in the paper.

Prayer to the AGW gods.

All that said, I don’t believe that warming is undetectable or that temperatures haven’t risen this century. I believe that CO2 helps warming along as the most basic physics proves. My objection has always been to the magnitude caused by man, the danger and the literally crazy “solutions”. Despite all of that, this temperature series is statistically speaking, the least impressive on the market. Hopefully, the group will address my confidence interval critiques, McIntyre’s valid breakpoint detection issues and a more in depth UHI study.

Holding of breath is not advised.


78 Responses to “Considered Critique of Berkeley Temp Series”

  1. Brian H said

    Targetted data set damage has been SOP for the AGW theorists and speculators from day one.

    BTW, your lukewarmism becomes you no more than it does other pusillanimous wafflers.

  2. kim said

    I’m about to wish CO2 had a decent warming effect. How otherwise could man actually engineer a warmer world so safely? And the need will be there to combat chilling, soon enough.
    ==============

  3. Jeff,

    I defer to your judgement on #1 and #3 and agree on #2. It is easy to dismiss UHI because of expectations that it should be ubiquitous in data from urban sites, when in reality that is not the true picture.

    There is no doubt that UHI exists, but it should not be expected to develop in a uniform manner and this makes detection more difficult. A disparity between urban temperature and that of surrounding areas can be easy to discern, but teasing out when the difference developed and how fast it grew, which is the real issue for warming, is not. I’m sure most people realise this if they have thought about it at all

    The BIG issue is those surrounding areas – and in particular the more contested warming at sites with small but growing populations (a la Roy Spencer’s analysis http://www.drroyspencer.com/2010/03/the-global-average-urban-heat-island-effect-in-2000-estimated-from-station-temperatures-and-population-density-data/), or that as a result of land use change. If these (as I am sure they do) cause warming in homogenisation and prevent better discernment of UHI, this is a big challenge to overcome – both in actual quantification and in overcoming the inertia of those who have no motivation to do so.

  4. Bad Andrew said

    “I believe that CO2 helps warming”

    Jeff,

    Helps? I agree with Brian H. Would you like some powdered sugar or syrup on that waffle?

    Sure it helps… except when it doesn’t.

    Andrew

  5. Jeff Id said

    #4,

    What waffle? I’m pretty clear in my views on CO2.

    Is AGW real – yes.

    How big is it – don’t know and neither does anyone else.

    Is it dangerous – Doubt it, but refer to 2

    How do we solve it – can’t.

    Simple

  6. Bad Andrew said

    Jeff,

    I don’t mean to nitpick, but there is a science of language too. When you add ‘help’ to your sentence, it sounds like you are talking to a grade school science class. And the ‘I believe’ part sounds more like the beginning of the Apostles Creed.

    The the more scientific statement ‘C02 makes it warmer’ is clearly not the whole story, because it doen’t ‘make it warmer’ when the temperature goes down.

    Why you cling to this AGW mess continues to escape me.

    Andrew

  7. kim said

    What else can we do but cling to the belief that the sun is steady? It is, but not in the way we’ve hoped.
    =============

  8. John Vetterling said

    I first posted this under the wrong thread:

    Something seems to be amiss in their scalpel technique. They say they look for abrupt jumps > 4-sigma. That indicates that the jumps should be on the order of 1/16K. If there is a record for each day that should equate to ~ 1 jump every 43 yrs. But they are reporting (if I read correctly) about 1 every 12 years.

    Am I missing something?

  9. Paul Matthews said

    The “CONUS temperature trend” – nice acronym.

  10. Jeff Id said

    “Why you cling to this AGW mess continues to escape me.”

    I don’t even know what you mean by that. Physics is physics, we don’t get a lot of choice in the matter.

  11. Bad Andrew said

    “Physics is physics”

    What else would it be?

    Not a very meaningful defense of AGW, either. Not much explanatory power.

    Andrew

  12. Genghis said

    About the slicing and dicing of the records, I am pretty sure that the UHI effect is pretty much a step change (or a series of step changes in response to the new variables). Just for kicks I put a thermometer (in a ventilated box) about 10′ from the south side of my house and about 100′ from a similar thermometer in an open field that I have been using for a long time.

    The thermostat close to the house immediately registered about 5 degrees (Fahrenheit) higher on average (max+min/2) in the winter and consistently maintains that temperature difference except when it is windy and/or overcast which it often is. In the winter when the snow melts from the south side of the house, the temperature differential increases dramatically with the thermometer next to the house sometimes registering 40 degrees higher. In the summer the two thermometers are often in agreement (the sun doesn’t reflect from the south side of the house). The long term temperature trend (just guessing based on daily observations) seems to be about the same, the only real difference is that the thermometer next to the house often seems to read about 5 degrees higher.

    So my dilemma is if I was a ‘climate scientist” how would I detect that a wall or some other physical change had occurred that affected the temperature reading? If I didn’t account for the new wall my temperature anomaly would gradually show up as a couple of degrees change over a period of time that might take years to average in.

    I don’t know that the two temperature records could be accurately spliced together with ‘adjustments’ and I am curious as to how long the 5 degree temperature increase takes to average into the long term trend of the temperature record.

  13. Jeff Id said

    #11, Ever read this:

    http://noconsensus.wordpress.com/2010/08/06/radiative-physics-simplified-ii/

    The first graph says it all (figure 2), the rest is important if the first graph doesn’t make sense.

    #12, I think you have several relevant points there which would be fun to discuss but I’m working and don’t have time. I’ll answer later.

  14. Bad Andrew said

    Jeff,

    Thanks for the link, but I don’t know what it is supposed to convince me of. It doesn’t explain how despite the alleged increase C02 in the atmosphere, the temperature still decreases. If there is a relationship between C02 and temperature, it ain’t direct.

    Andrew

  15. Genghis said

    Bad Andrew “If there is a relationship between C02 and temperature, it ain’t direct.”

    The link clearly showed that there is a direct relationship between temperature and CO2, if everything else is carefully constrained. We all know that everything is NOT carefully constrained in the real world. We aren’t even sure of the signs or magnitudes of ‘everything else,’ that is where the debate is. Warmers simply assume that ‘everything else’ (primarily water vapor) except for aerosols is a positive influence. Some of us think that there are things like clouds, convective cells, albedo changes, etc. that may in fact be potential negative drivers.

    What the ‘Climate Scientists’ are trying to do is tease out the “CO2 signal” from the temperature record with complex computer models. The fact that their computer models failed to predict the current lack of warming and ENSO variations, clearly falsifies the current models. Of course it isn’t really the modelers faults that the models fail, we don’t know all the basic parameters to feed into to the model. That is what this thread is about, one of the important parameters is the temperature record and not only does it measure the wrong thing (energy vs temp) it isn’t even a very good temperature record.

  16. Jeff Id said

    #14, It isn’t my job to work it out for you, the basic physics are simple and proven yet there are huge energy sinks and feedbacks. If in 20 years someone proved that CO2 caused nearly zero net warming because feedback was strongly negative, I would be surprised but it isn’t impossible. Everything seems to line up to mild warming except for the models and some land temp series. I’m complaining here about the land temp series which is only 30% of the global temp record.

    Genghis,

    Regarding the detection of steps in your thermometer record, it would be nearly impossible. Imagine though that you had a hundred thermometers nearby the 100′ one. Then you could detect the individual by your house as a statistical anomaly which is what BEST tries to do. De-weight that record and you would get a less contaminated signal. In climate science, instead of a lot of good records and one bad one, a large number of temp stations are polluted. The polluted stations in BEST get sliced and realigned based on sudden steps relative to a regional average. Unfortunately there are cases where the step may be important to keep. What if a town grew and the thermometer station was repeatedly located in less urban locations moving from the center to the edge of town. Each move produces a downstep which BEST slices and reattaches the data producing an upward trend which grows between steps as the surroundings urbanize. In reality, the steps were intentional corrections by the station operators not to induce urban warming, in practice, BEST has just undone their work to create an increased uptrend. McIntyre made this example in a comment.

  17. “Each move produces a downstep which BEST slices and reattaches the data producing an upward trend which grows between steps as the surroundings urbanize.”

    I don’t think that’s what they do, if you’re referring to the scalpel. I think they slice, but don’t reattach. They just treat them as separate stations. The criticism of that is that they lose a lot of information, not bias.

  18. Jeff Id said

    Nick,

    If you imagine the sections re-attaching, I beleive they would be attached to the mean (generally speaking). If so, then the step is removed despite the separate treatment. Is there some instance you can describe where the fidelity of the station I described would retain the correct steps?

  19. Bad Andrew said

    “We all know that everything is NOT carefully constrained”

    Genghis,

    That why you can’t say “C02 causes warming” and still be correct. Because it can’t, if it gets colder. You’d have to say “C02 may contribute to warming, if it gets warmer.”
    It’s obviously not contributing to any warming if there is no warming, like when the squiggly line goes down.

    Andrew

  20. Jeff, I don’t imagine them reattaching at all. They are just treated as independent stations. You had a station in the city. It went for a while and closed. Then one started in the suburbs, went for a while and closed. Etc. That’s how they are treated after scalpeling. . They aren’t linked.

  21. Jeff Id said

    Nick,

    Seriously, is there a way that you can imagine the series I described having the proper amount of effect on the average? All the segments I described are UHI contaminated so how would these be properly treated?

  22. Politicized science and exponentially improving abilities to compute and extrapolate scientific models to predict the future seem to have almost completely destroyed the reliability of:

    1. The UN’s global climate model,
    2. The Bilderberg SSM solar model [a], and
    3. The Greenspan/Bernanke economic model for the Federal Reserve [b]

    [a] Solar Physics 3, 5-25 (1968)]

    http://adsabs.harvard.edu/full/1968SoPh….3….5G

    [b] The Federal Reserve

    http://www.save-a-patriot.org/files/view/whofed.html

    Moral: The reliability of computer model projections are inversely proportional to a.) political pressure and b.) number of adjustable variables.

  23. Jeff,
    UHI is a separate issue, and you can’t do anything with it without bringing in information from other sites.

    But the scalpel idea is simple. You have a sequence that is initially presented as a single site, but has a jump. One possibility is a station move. Well, if that is what happened, it does mean that you close one station and open another in a different place. If they were far apart, you’d regard these as separate events, and there wouldn’t be an issue. If they are close, then you can if you like try to exploit the fact that one is a continuation of the other in some way. But you don’t have to. You can still suppose the other is just an independent new station. I think that is what the scalpel does.

  24. Jeff Id said

    Nick,

    UHI is the point I am making regarding the scalpel algorithm. Each time a station is moved, a preferred out of the way location is found. It is reasonable to assume that the station has been moved because it was in the way of something needing the land meaning that growth was happening around it before the move. If this case is common than chopping of the steps may be the worst thing you could do. I didn’t consider this myself previously but now if you find a step in the station, chucking all the data may be correct, keeping the data intact may be correct, steps up may be bad, but steps down are more likely to be good. The method is more likely to detect a downstep just because of the uptrend. So the whole thing is a mess. Perhaps the best reconstruction you could do would be to chuck anything suspicious for steps or urbanization but without a true metadata documented reason, that could bias the record the other way.

  25. j ferguson said

    Jeff, does Best inventory the direction and magnitude of the step changes that led them to end the first record and create a new one? I take it that a record discontinuity was enough for the scalpel, but it might be interesting to characterize the amount of any trend change in the scalpeled tail. I think I understand that they down-rate the post scalpeled tail, but why?

  26. NIck, JeffID is correct. The scapel is to calibrate the splices wrt the mean. His comment is a correct and valid observation.

    Though the first step is not the mean but somethig else. The comparison eventually is to a mean just as GISS, etc.

  27. cce said

    The scalpel is not intended to address urban warming. It’s intended to address discontinuities like station moves, instrument changes, and other microsite issues (assuming they’re detectible). You address UHI induced warming (not UHI which is not questioned) by identifying rural stations and comparing them to non-rural stations. This requires good urbanity proxies and accurate location information, both of which are hard to come by, so it requires very careful analysis. A station that is moving around inside of an urban area would not be considered rural and would be excluded.

  28. Jeff Id said

    Cce,

    If you re-read what I wrote, I am showing an example where due to UHI, the method addressing instrument moves exacerbates the problem. Are there instances where a station move shifted it to a worse location? Probably a lot of them but typically you would expect the operators to choose something out of the way.

  29. I put the following comment at Climate Audit
    Has not the UHI effect been long recognised? Torok et al in Aust. Met.Mag 2001 in the paper “Urban heat island features of southeast Australian towns” gives the relation delta T=1.42log(population) -2.09. This was followed up by Morris et al with a travers of Melbourne Aust. Prof Ole Humlum (see http://www.climate4you.com/index.htm) has made traverses of Oslo, Spitzbergen and other towns in 2007. I have a outside thermometer attached to my Subaru. Just about everyday I note the temperature difference from the CBD of the local town (population about 30,000) compared to my home on acreage at the semi-rural outskirts (forests, some paddocks, a strawberry farm etc). I note the temperature 1 to 2C higher in the CBD at all hours (day, dusk, night, dawn). You can note the temperature effect driving along the highway. The capital city can be upto 4C higher than the rural surrounds. It is not only sloppy that CRU and others deny UHI. One could regard it as scientific fraud.
    Best and others should not only look at measured data of rural & urban related stations but ago out and do some traverses to calibrated the stations.

  30. cce said

    Jeff,

    The scalpel is not intended to correct for urbanization. Splitting off rural does this. If the rural sites warm at close to the same rate as all sites, then the cases that you imagine are either not very common or make little difference. The key, of course, is good urbanity proxies and and precise location information. This is not to say that BEST is the last word on this as there is certainly room for improvement. But the idea is correct: use the scalpel to address discontinuities that would otherwise require good metadata (which don’t exist). Use station location to address urban warming.

  31. cce said

    Cementafriend,

    CRU, NASA, NOAA, BEST, etc do not “deny UHI.” No one “denies UHI.” The question is whether urbanization biases global temperature analyses, and if so, how to correct for it. These are different things.

  32. Jeff Id said

    Cce,

    I don’t share your opinion about scalpel. I think it is a good concept which in practice is full of complications that can induce spurious trends. There is no reason to believe rural would produce the same trend as urban stations and even the detection bias could add more trend. It is nearly impossible to imagine how the method doesn’t exaggerate average trends – at least a little.

  33. Hm, either I’ve missed something, or you’ve missed a whole paper published by the BEST team: http://berkeleyearth.org/Resources/Berkeley_Earth_UHI.pdf

    “Time series of the Earth’s average land temperature are estimated using the Berkeley Earth methodology applied to the full dataset and the rural subset; the difference of these shows a slight negative slope over the period 1950 to 2010, with a slope of -0.19°C ± 0.19 / 100yr (95% confidence), opposite in sign to that expected if the urban heat island effect was adding anomalous warming to the record. The small size, and its negative sign, supports the key conclusion of prior groups that urban warming does not unduly bias estimates of recent global temperature change.”

    Thus #2 is simply irrelevant.

  34. Jeff Id said

    Grzegorz Staniak,

    The non-detection of UHI is NOT a sign of a good result, in fact it is the opposite. The method of the UHI paper is non-functional.

    • @Jeff Id

      Why is the UHI-detection method “non-functional”? And how could anyone a priori know that “non-detection of UHI is NOT a sign of a good result”? This loooks suspiciously circular to me (“good result is a detection of UHI”).

  35. Jeff Id said

    Grzegorz Staniak,

    Fortunately there is other evidence like the difference between sat and surface records. The known delta from rural to urban that couldn’t have been there a hundred years ago, the surfacestation project, the discrepency in trend between surface and ocean etc…

    • @Jeff Id

      The difference between the sat and surface records in itself doesn’t tell anything. It’s not that satellite temperature series are somehow by definition superior or referential. Quite to the contrary — the results for troposphere are by definition contaminated by a cooling trend in the stratosphere, and the process of arriving at a temperature series is heavily dependent on the choice of algorithms (e.g. homogenization of data from various satellites) and corrections: for the effects of the diurnal cycle drift, for the orbital decay etc. Corrections applied to UAH were able to increase the linear trend by as much as 40%, as you may recall. If you have a look at these:

      http://www.skepticalscience.com/Primer-Tropospheric-temperature-measurement-Satellite.html
      http://www.skepticalscience.com/Eschenbach_satellite_part.html

      and the references therein, you’ll see that alternative analyses of the satellite data can show trends considerably higher than even the surface series, up to 0.2 C per decade.

      The discrepancy in trend between surface and ocean is simple physics and I’m not sure why you’re mentioning it. Put a metal bar and a glass of water on a stove and see which will get hot sooner. As long as the heat capacity of the oceans is not exhausted, their temperatures will not rise in step with the land surface — and the capacity is enormous:

      http://www.skepticalscience.com/graphics.php?g=46
      http://www.skepticalscience.com/Ocean-Heat-Content-And-The-Importance-Of-The-Deep-Ocean.html

      Now, the surfacestations.org project has so far only confirmed that the UHI effect is insignificant: John de Vliet’s analysis at the milestone of 30% of classified US stations, in 2007, found no effect to talk about. Since then Anthony Watts had become a bit reluctant to release results — to the point of removing the data from the surfacestations.org site — but as long as they were available, independent analyses of them (e.g. at the 70% milestone) again confirmed no significant UHI effect:

      http://www.ncdc.noaa.gov/oa/about/response-v2.pdf
      http://www1.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/menne-etal2010.pdf

      So far we have no publication on surface stations from Watts other than on the blog and the surfacestations site, but quite recently, after submitting his own publication and after the pre-release of the BEST results, Anthony Watts himself had this much to say:

      http://wattsupwiththat.files.wordpress.com/2011/03/response_to_muller_testimony.pdf

      Temperature trend estimates vary according to site classification, with poor siting leading to an overestimate of minimum temperature trends and an underestimate of maximum temperature trends, resulting in particular in a substantial difference in estimates of the diurnal temperature range trends. The opposite-signed differences of maximum and minimum temperature trends are similar in magnitude, so that the overall mean temperature trends are nearly identical across site classifications.

      In other words, the “bad” stations underrate the maximum temps and overrate the minimum temps, so that the two biases cancel each other and general trends are the same as in “good” stations. I’d really expect no bombshells from Surface Stations if I were you.

      I must say I don’t really understand this ado about the UHI effect anyway. It’s been known for decades, estimated and accounted for. If you have a look at how the GISTEMP series is constructed:

      http://data.giss.nasa.gov/gistemp/

      you’ll find the following under the “Current Analysis Method” heading:

      The GHCN/USHCN/SCAR data are modified in two steps to obtain station data from which our tables, graphs, and maps are constructed. In step 1, if there are multiple records at a given location, these are combined into one record; in step 2, the urban and peri-urban (i.e., other than rural) stations are adjusted so that their long-term trend matches that of the mean of neighboring rural stations. Urban stations without nearby rural stations are dropped.

      In other words, the long-term trends in GISTEMP are actually based only on the “good” stations data.

  36. Howard said

    Grzegorz Staniak makes a great point. UHI may not be a big deal in the surface temperature record and some other mechanism might be responsible for the discrepancies Jeff ID cites. For example, lets say that data from the surfacestation project properly crunched into temp records indicates that most sat and surface difference are due to micro-site problems… That is not UHI. Perhaps UHI is masked by rural landuse changes (or macro-site problems) that give false heating trends in the entire network. Neither of these two potential causes of false positive temperature trends have been quantified.

    The difference between the ocean and the surface trends are probably dominated by differences in heat capacity and advection.

    BEST is the end of the beginning that has confirmed that simple statistical models are not sufficient to sort out the discrepancies. Now the hard work of identifying the problems at the level of each station, each station move, each thermometer, each area, region and continent needs to be sorted.

  37. @Howard

    Actually, the trends are much more important than absolute precision of measurements. Just as for a hospital patient — if you read a series of rising temperatures measured orally, you don’t delibarate on the divergence of 2 C from the rectal readings, and you don’t sit down to design a perfect system of measuring the average body temperature. It’s enough to know that the oral reading is representative, and that however different other readings may be, if they show a similar trend, you have evidence of a fever.

    Here’s a nice explanation of the whole process:
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_1A.html
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_1B.html
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_2A.html
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_2B.html

  38. Jeff Id said

    Grzegorz Staniak,

    The difference between the sat and surface records in itself doesn’t tell anything. It’s not that satellite temperature series are somehow by definition superior or referential. Quite to the contrary — the results for troposphere are by definition contaminated by a cooling trend in the stratosphere, and the process of arriving at a temperature series is heavily dependent on the choice of algorithms (e.g. homogenization of data from various satellites) and corrections: for the effects of the diurnal cycle drift, for the orbital decay etc

    This is not correct – at all. First, I strongly recommend that you get your information from the papers rather than an advocacy blog disguised as science. I can’t even read the thing. Our comparison is between lower tropospheric and ground station data rather than tropospheric trends. You can see from the vertical weighting plots that there is minimal weight in the stratosphere and the measurements are verified in numerous papers using radiosonde. In addition, the diurnal drift corrections were planned to be in opposite directions over time so that two satellites drifting oppositely were used to minimize any trend effects. Recently UAH had been using the AQUOS satellite until it kicked off, and it had station keeping which eliminated the correction entirely. I believe it will be back on a station keeping sat next year. Finally, if you state that UAH and RSS had a correction done to it in the past as reason to not trust them, then you are not understanding the history of satellite measurement well. The correction was actually quite small, openly discussed and repaired and we still have a far lower trend than ground.

    Temperature trend estimates vary according to site classification, with poor siting leading to an overestimate of minimum temperature trends and an underestimate of maximum temperature trends, resulting in particular in a substantial difference in estimates of the diurnal temperature range trends. The opposite-signed differences of maximum and minimum temperature trends are similar in magnitude, so that the overall mean temperature trends are nearly identical across site classifications.

    Now that we’ve clarified that there is a true and valid discrepancy in satellite vs ground whcih cannot be waived away by an advocacy blog, ask yourself if there is a possibility that classifications are not adequately done. This is what Anthony Watts surfacestation project identified as a problem.

    Regarding surfacestations, their conclusions were not a non-problem but didn’t reveal a huge discrepency:

    and the differences in trends compared to CRN 1&2 stations were statistically significant for all groups of poorly sited stations except for the CRN 5 stations alone. The magnitudes of the significant trend differences exceeded 0.1°C/decade for the period 1979-2008 and, for minimum temperatures, 0.7°C per century for the period 1895-2009.

    I’ve been asking a different question than you. I’ve been asking why didn’t they detect a stronger signal when we have known large UHI discrepancies that didn’t exist 50 years ago?

    If you have a look at how the GISTEMP series is constructed:

    I have spent a great deal of time on temperature series and have even created a global average which is not substantially different than the others. Until you look at the data, you really don’t get a true feel for the size of the problems. So we have GISS methods and we have BEST methods. I wonder if you believe that either of these methods would detect a new building or concrete pad being constructed directly next to an existing station in the last 100 years? What would that do to temperature at that station?

    I wonder how people are so certain that microclimate UHI type contamination isn’t a bigger effect. My opinion is that the satellite and radiosonde data are proof that the problem exists and good evidence of the true magnitude.

    Now regarding your point on ocean vs land trends being expected. Excepting the water/metal bar experiment which doesn’t make sense in this context to me, I agree with you completely. I’m sorry for the lack of clarity in my comment. My point is not that there shouldn’t be some difference in a non-static temperature globe, but rather that the difference in the trend two is not adequately explained.

  39. Carrick said

    Jeff ID:

    Now that we’ve clarified that there is a true and valid discrepancy in satellite vs ground which cannot be waived away by an advocacy blog

    The abject stupidity of blogs like SkS is they gloss over real science simply because the difference doesn’t fit into their pin-headed narratives. And then its denizens happily swallows their koolaid (taught to “group think”) rather than being encouraged to think individually and question. Blog authors rarely admit to substantive error, rather they simply edit away embarrassing errors, leaving in comments from critics while changing their responses to make it look like the critic never even read the original article.

    You have to correctly account for the fact that surface temperature measurements are made in the surface boundary layer whereas satellite measurements are made at an elevation well above the boundary layer.

    Any by “correctly account” I mean something different than see whether a particular GCM is able to account for the differences. [Hint: It can’t. It doesn’t have a surface boundary layer in the model at all. So direct comparison of GCMs with surface data is impossible without a model connecting, probably something like the ECMWF reconstruction is as close as we’ve come.

    The atmosphere gets up to a lot of radical sh*t in the first few meters above the surface. Just compare daytime to nighttime measurements. (Daytime is about 2-pm if I remember.)

    A few comments:

    1) Note that daytime temperatures have increased temperature gradients near the surface, but the lapse rate is typically negative. At night time near the surface it can be large and negative (the ground cools more rapidly than the air above it due to radiative cooling, just as it is heated more rapidly in the day time.)

    2) The amount of the temperature change depends on surface coverage as well as humidity and wind speed, and depends on time of day.. There are models that allow you to calculate what it would typically look like in Alaska with snow cover versus at an airport in Florida versus in a deep canopy in Ecuador.

    3) Over ocean there is relatively difference between day and night time boundary layers, and it is typically always negative, and the temperature trend is nearly constant with latitude, due probably to ocean currents.

    4) land temperature trend increases with latitude. Since it is a maximum at the more northern land latitudes, this is not consistent with UHI contributing a substantial effect to land temperatures. (If UHI were a dominant contribution it should be peaked in regions with the most rapid urbanizations. China and India are near the equators, most of the “western” civilizations and Japan’s urbanization occurred below the 45°N parallel.)

  40. Carrick said

    Corrected comparison of day to nighttime temperature trends.

  41. Carrick said

    And make that below the 55°N parallel in Europe. In the US it’s below the 45°N parallel.

    It would be interesting to look at the population density distribution too. I don’t think it’s going to have much to do with the latitudinal effect on land temperature trend I showed in the figure above though.

  42. Howard said

    Grzegorz Staniak

    My absolute body temperature is much more important than having an accurate trend to a couple tenths per decade or per hour for that matter. Are you just repeating some pothead pipedream analogy that sounded good at the spleefical science blog?

    I hope that my previous compliment encouraged you to come out of the closet, so to speak.

  43. Howard said

    Carrick:

    4) land temperature trend increases with latitude. Since it is a maximum at the more northern land latitudes, this is not consistent with UHI contributing a substantial effect to land temperatures. (If UHI were a dominant contribution it should be peaked in regions with the most rapid urbanizations. China and India are near the equators, most of the “western” civilizations and Japan’s urbanization occurred below the 45°N parallel.)

    Isn’t the high latitude north warming (and a blip in the south) strongly influenced by humidity? Conceptually, more heat on the equator goes into doing work that is not reflected in dry-bulb temperature… or am I in left field.

  44. Carrick said

    Howard:

    Isn’t the high latitude north warming (and a blip in the south) strongly influenced by humidity?

    I think that is certainly part of it (less heat loss at night). There is also an effect due to change in surface albedo resulting to changing patterns in snow coverage (so this is more of a change in seasonal patterns.)

    My guess is if you really wanted to separate out the various boundary layer effects, you’d look at tmin, (true) tavg and tmax trends separately, as well as do the comparison of the trends by seasons.

    Of course that’s been done to some extent. With the data I have access to, I could break it out by season and latitude. If somebody knows of a break out by tmin and tmax (that is on e.g. 5°x5° grids) that would be very useful.

  45. Jeff Id said

    Carrick,

    Regarding your point number 4 above, I wonder if population density is more of a contributor to urban effrects than industrialization. For instance, in India and China poverty would prevent them from placing instruments in the countryside and the cities were already heavily populated. In Japan, I don’t know if it was possible to locate an instrument in a truly rural area wheras areas of Europe and northern US and Canada might have seen a far greater effect as we encroached on rural areas. It’s just a concept, but to me, the huge hemispheric trend discrepancy is another strong piece of evidence that something in addition to real warming is going on.

    Yes we know of the ocean area explanation but the trends are strongly different as your plot shows.

  46. Carrick said

    Jeff ID:

    Regarding your point number 4 above, I wonder if population density is more of a contributor to urban effrects than industrialization. For instance, in India and China poverty would prevent them from placing instruments in the countryside and the cities were already heavily populated

    That would indicate you’d get a bigger UHI effect there, not smaller, right?

    Yes we know of the ocean area explanation but the trends are strongly different as your plot shows.

    I think much of the difference between ocean and land the oceans “mix” and have a high heat capacity, and the day to night temperature (sans weather fronts) doesn’t change that much.

    It would be very interesting to look at just tmax versus latitude, land versus ocean.

    I predict you won’t see as big of a difference.

  47. Carrick said

    This is probably a good place to start.

    “A stable boundary layer perspective on global temperature trends” by R.T. McNider, J.R. Christy, A. Biazar.

    . One of the most significant signals in the thermometer-observed temperature record since 1900 is the decrease in the diurnal temperature range over land, largely due to warming of the minimum temperatures.

  48. Carrick said

    The empirical question of effect of latitude & land versus sea on temperature trend (min,max,avg) can be answered I think without having to understand the boundary layer physics behind it. Merely knowing there is a theoretical foundation for such an exploration is I think sufficient for purpose of testing.

    If you want to claim that most of the observed warming is due to microsite issues related to UHI (which is very different than just saying changes in land usage acting as a forcing for climate change), I think it’s a very long, steep climb at this point.

    I’m sympathetic to the idea it is important and should be quantified, and expect if mishandled it could lead to a systematic temperature bias (which you have to sniff out using a Monte Carlo based approach, IMO). But I don’t see how it can be more than 10% or so of the total observed trend over land…. in other words, Phil Jones assumption you can neglect is probably not only validated, it probably is a safer course of action to take than to use some ad hoc method that may correct microsite issues in some cases, but make it worse in others.

    [This comment is related to the SkS group's love affair with GISTEMP. They seem to equate more complicated algorithms with necessarily improved ones. That is just so wrong on so many levels.]

  49. Jeff Id said

    I wouldn’t claim that ‘most’ of any warming is microsite. I would claim that the Satellite record is the more accurate one and the difference is mostly UHI. We know the methods used don’t have any sensitivity to the construction of roads and buildings near temp stations. We also know that most station moves are driven by the need to use the land for more construction. My guess is that the effect is seriously underestimated.

  50. Jeff Id said

    Again though, the models are running hot in comparison to what I think are overheated warming curves. The differences should not be ignored by scientists as easily as they seem to be. If AGW were a non-political issue, the models would have already been changed.

  51. Kenneth Fritsch said

    I think that the UHI construct is overly generalized by both skeptics and those of the consensus on AGW. It is troubling for me to see the GISS group under Hansen using satellite night lights and population to quantify a UHI effect and to depend solely on rural stations (as classified by GISS) for trends. Some skeptics seem to argue that since the UHI effect can result in higher temperature heat islands generally that that must translate to a significant bias in the calculated temperature trends.

    I would say that, until some comprehensive testing is done on individuals sites and the effect of the micro site climate on temperature measurements, we could be assuming effects far too generally. The CRN rating of stations carried out by the Watts team would merely be a start in a proper analysis. The analyses of CRN rated stations on temperature trends in the Fall and Menne papers used too short of a time period and were not sufficiently detailed. A major problem in these types of analyses will always be the lack of meta data to cover the timing and extent of the station changes, but at least we could put some boundaries on the problem if the station change effects were better quantified. My guess would be that we might see more uncertainty in the temperature trends without necessarily finding different mean trend values.

    My major point here would be that micro sites lying within a general cool or warm zone could have temperatures significantly different than the generalized area. In Chicago over past decades we had an official station change from Midway to O’Hare airport. Midway’s temperatures may have had some UHI effect since it was located well within the city proper, but was probably most affected by its proximity to the lake and thus its micro/macro climate was significantly moderated by Lake Michigan. While O’Hare is officially within the Chicago city limits it is actually out in the suburbs and its temperature have much less moderation from the lake than is the case for Midway.

    As an aside, I was pondering how climates were modeled for the BEST Monte Carlo on station variability. Do not climate models lack the resolution to model the local climates required to emulate station climates?

    Finally, I must repeat here that I am having great fun analyzing breakpoint data for the USHCN temperature series and thinking about what the results portend for localized climates.

  52. Carrick said

    Jeff ID:

    I would claim that the Satellite record is the more accurate one and the difference is mostly UHI

    I am pretty sure that 1) that’s not consistent with the data (while is the trend the largest in a systematic way at the most northern climates, I think that blows this hypothesis out of the water, does it not?) and 2) satellite may be an accurate measurement of something, but that something is certainly not temperature measured 2-m off the surface of the ground.

  53. David S said

    If the papers suggesting a broad logarithmic relationship between population and UHI are correct in principle, then the first few thousand inhabitants in a previously rural area will warm the observed local temperature up more than the next million in an urban area. So wherever rural sites are swallowed up by sprawl we could expect a significantly enhanced trend, and it would only need to be a minority of such cases to distort the data set as a whole. This would go some way to explaining why the trends within the global averages constructed by Zeke, BEST and others show little difference between urban and rural sites.
    Carrick can I ask how rural are sites like Barrow in the far North, and don’t they play a significant role in the gridded reconstructions because of the relatively scarce coverage of stations in high latitude areas? Here is a paper that I turned up that someone at University of Alaska did back in 1991 where the issues seem to have been better understood than they are now!
    http://climate.gi.alaska.edu/bowling/AKchange.html

  54. Carrick said

    David S, here is the problem:

    The trend from equator to most northern extent is nearly monotonic:

    Figure.

    My suspicion is it would take a remarkable amount of tuning in a model of population growth density, economic growth and UHI effect to replicate this—a model that posits the dominant latitude effect comes from UHI is pretty much blow out of the water, IMO.

    Finally this type of effect is largely expected in the event of global warming, and is consistent with other observations such as loss of permanent land ice (and associated decrease in surface albedo).

    So one model needs a huge amount of tuning and makes the GCMs look elegant in comparison, and the other is a natural consequence of the science of global warming.

    Which of these is the more plausible in the sense of Ocaam’s Razor?

    This doesn’t mean I don’t think you can absolutely neglect UHI, nor do I think it is an “undetectable signal.”—plausibly it could represent as much as 10% of the surface warming at some latitudes. But I think at that level, there are a host of other site related issues of that same order: site moves, changes in instrumentation, measurement protocol, etc. that need to be addressed.

    I happen to agree with BEST that the most explanatory variable for geographical variation in surface land temperature trend comes from latitude, secondly from elevation of the measurement, but this is a just proxy for whether it is coastal/inland if you think about it—and we already know that marine boundary layer is much less volatile than the surface boundary layer.

  55. Carrick said

    Minor correction to my comment “… and we already know that marine boundary layer is much less volatile than the [land] surface boundary layer”

  56. diogenes said

    I wish people would stop invoking William of Occam..those scholastic philosophers wee rather more rigorous in their approaches than bloggers

  57. Carrick said

    Diogenes:

    I wish people would stop invoking William of Occam

    Not gonna happen, unless you want to rename the test something else.

    It’s become a formalized aspect of statistical analysis… how many parameters you need versus how explanatory the model is. See e.g. AIC as an example of such a metric.

  58. Carrick said

    Fixed AIC link.

  59. Jeff Id said

    Carrick 54,

    Just back to the computer. I’m not understanding- really. Why would greater trends in the north not represent greater UHI in the north? The north has experienced the greatest development and kept the most records (more chances for moves by encroachment too).

    On point 2, SteveM’s recent post shows that models predict satellite trends higher than ground not the other way around. It is a different measurement but it is also lower troposphere.

  60. Carrick said

    Jeff ID:

    Why would greater trends in the north not represent greater UHI in the north?

    Theoretically it could, but I think you’ll find that population growth doesn’t follow this simple pattern: Most recent development is in the third world—think towards the equator, not towards the far north. My expectation is if you tried to build a model to predict which latitudes would have UHI enhancements, it wouldn’t monotonically vary in the way that’s observed.

    As I suggested above, it would take quite a balancing of population & economic growth and site placement to end up with something this systematic.

    However, the pattern is fully consistent with polar amplification expected global warming from any source (natural or manmade). So on the one hand you have rely on a pattern of change in UHI that doesn’t really track with growth of population centers, and on the other, you have a trend that seems consistent with the expected physics.

    Obviously I’m going with the latter choice here. If somebody wants to try and model UHI and show it can replicate this pattern that would be interesting, but I’d bet quatloos it would require a lot of tuning to get it to work out.

    On point 2, SteveM’s recent post shows that models predict satellite trends higher than ground not the other way around

    The models don’t even have a surface boundary layer, so I don’t really see how you can apply land surface temperature measurements to their predictions…. not without some type of modeling to relate the GCM output to measured surface temperature.

  61. Brian H said

    There’s another issue re station siting and UHI bias: modernization. When manually recorded, stations could be at some distance from the nearest building. But when cable has to be run, it can’t go under pavement, etc. very easily, so the stations were relocated close to the recording building/equipment. That often eliminated the “buffer zone” and brought the station up next to the A/C outlets, paved surfaces, building walls, etc., etc.

    So electronic convenience sabotaged the integrity of the station and its isolation from obvious heat contamination.

  62. Howard said

    Carrick:

    Thanks for the link to the boundary layer paper.

    “The temperature at night at shelter height is a result of competition
    between thermal stability and mechanical shear. If stability wins then
    turbulence is suppressed and the cooling surface becomes cut-off from the
    warmer air aloft, which leads to sharp decay in surface air temperature. If
    shear wins, then turbulence is maintained and warmer air from aloft is
    continually mixed to the surface, which leads to significantly lower
    cooling rates and warmer temperatures. This warming occurs due to a
    redistribution of heat.”

    This concept from the abstract reminded me of vortex generators that are used on airfoils to shrink the boundary layer. I imagine that new buildings are like adding vortex generators to the land surface with the potential to raise local night-time temperatures at lower relative wind speed. Conversely, the clearing of land for agriculture would remove vortex generators (grading land and removing trees) with the potential to decrease local night-time temperatures at higher relative wind speed.

    Perhaps both rural and urban anthropogenic land-use changes which alter “natural” boundary layer dynamics are responsible for much of the discrepancy between sat and surface temperature trends . The rural component of this potential false warming measurement is critical because The consensus on UHI is based entirely on the rising rural temperature trend. In other words, if all regional, local and micro-site effects (boundary layer, asphalt, concrete, irrigation, etc.) can be accounted for as insignificant factors in the rural station data (Rural Heat Zones or RHZ), then UHI is not a factor for the surface record. I suspect that the surface record difference from the satellite record is at minimum 50% due to a combination of UHI plus RHZ. The other 50% is likely due to a combination of satellite temperature calculation errors and how a warming planet distributes heat.

  63. DeWitt Payne said

    Re: Carrick (Nov 10 21:57),

    The models don’t even have a surface boundary layer, so I don’t really see how you can apply land surface temperature measurements to their predictions…. not without some type of modeling to relate the GCM output to measured surface temperature.

    They don’t have a boundary layer, but they should have a surface temperature, or at least an effective radiating temperature. It would be a little hard to model radiative transfer if you didn’t know how much energy was being radiated by the ground.

  64. @Jeff Id

    What is “not correct”? I’m not arguing that tropo- and stratosphere trends should be compared with the surface record. What I’m saying is that satellite temps are not somehow carved in stone and by definition superior to the surface record. You say: “fortunately there is other evidence like the difference between sat and surface records”. I say: “in itself it’s NOT evidence of anything” — in particular this fact by itself is NOT an indication that there’s something wrong with the surface temperature series. Can you deny that since version A the UAH satellite record decadal trend has been corrected +0.7 C? If Skeptical Science as such is not skeptical enough for you — judging by your blogroll, you have a strong bias towards the denialist positions — then consult the papers directly:
    http://www.ncdc.noaa.gov/temp-and-precip/msu/nature02524-UW-MSU.pdf
    http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.143.4699&rep=rep1&type=pdf
    http://www.agu.org/journals/ABS/2006/2005JD006798.shtml

    Sorry, but UAH/RSS are neither beacons of stability, nor the only valid way to look at satellite data.

    And Jeff, I’m sorry, but if you deal with temperature series and still ask the question “why didn’t they see stronger UHI effect in comparison with 50 years ago?” then you’re just trolling. I gave you the URL, I’m sure you can find more info yourself: it’s because the UHI effect has been known, explained, estimated and accounted for. The long term trend in e.g. GISTEMP is based exclusively on “good” (rural) stations, since all urban and peri-urban station readings are adjusted to the mean of neighbouring “good” stations, or simply dropped when none such station readings are available.

  65. Jeff Id said

    “it’s because the UHI effect has been known, explained, estimated and accounted for. ”

    You sound pretty sure of yourself. I’ve read the papers and am less so.

  66. Carrick said

    DeWitt:

    They don’t have a boundary layer, but they should have a surface temperature, or at least an effective radiating temperature. It would be a little hard to model radiative transfer if you didn’t know how much energy was being radiated by the ground

    True, but good luck predicting how much surface amplification you get without a boundary layer model (see my reference above for context).

  67. Carrick said

    Howard:

    This concept from the abstract reminded me of vortex generators that are used on airfoils to shrink the boundary layer. I imagine that new buildings are like adding vortex generators to the land surface with the potential to raise local night-time temperatures at lower relative wind speed

    That’s a clever idea. Even if not for GCM, it might be useful in agriculture (prevent those early spring freezes).

    On a related topic, one of the more initially mystifying things that happened when I started collecting nocturnal met data (this is my 10-m tower</a) was that when you get wind bursts at night it often was associated with warming events near the ground.

    We eventually worked out that what was happening was during big wind gusts, warm air from higher elevations were getting circulated down to the surface (in this particular data set, it was on average about 10°C warmer 100-m above the surface about 2 hours after sunset… we had a tethered balloon to gives us measurements at higher elevations).

    Almost every time we've gone out at night to collect data, we've encountered something unique about the meteorology. It's a bit mystifying to me how you relate the output of global climate models to phenomena that are this finally tuned.

    Oddly, daytime btw is much more predictable, because the presence of turbulent driven convection. The nocturnal layer is so unpredictable precisely because it is truly a metastable condition—not "truly stable" because that stable layer is a fiction invented for meteorology text books. In the real world, it gets up to stuff.

  68. Kenneth Fritsch said

    I would guess that most of us agree that there is a UHI effect and that temperatures within a large urban area can be significantly warmer than the surrounding less urban areas. The issue in contention is what effect UHI has on long term temperature trends – and as I see it that could be minimal. I am more interested in the micro climate effects and how changes in the micro climate can effect trends or the CIs for trends. I personally conjecture that micro climate can overwhelm generalized urban versus rural considerations.

    Maybe we should discuss the absolute UHI effect and what that portends for the effects of GW on humans and their abilities to adapt to changing temperatures. Do we know whether the lack of evidence for trends in urban areas is due to changes in the siting of thermometers. Such a discussion should also bring into view the detrimental/beneficial effects of AGW and perhaps separate those issues that deal with regional to micro climate changes and human’s abilities to adapt to those changes versus those that might be related global temperature changes like for example rising sea levels, melting glaciers, drought floods, violent storms, etc. If we are supposed to be scared it might be best to know what we should be afraid.

  69. Carrick said

    Kenneth Fritsch:

    I personally conjecture that micro climate can overwhelm generalized urban versus rural considerations.

    Yep. That’s a point I was making on Lucia’s blog today also.

    UHI can amount to 2°C total. Microsite effects can be much larger.

  70. Carrick said

    Christy gives a discussion of the surface station project that includes the impact of environment on tmin. Menne has comments at the the end that seem to demonstrate that these comments flew over his head.

  71. @Jeff Id

    Care to elaborate your doubts? Or at least point to those papers that you read that make you less sure about the significance of UHI? Because, you know, at the moment I have your “I’m less sure” on the one hand, and e.g. this list:

    http://agwobserver.wordpress.com/2009/12/27/papers-on-urban-heat-island/

    on the other. Guess what looks more convincing.

  72. Carrick said

    Interesting selection of UHI papers.

    Schmidt 2009…. Using a GCM, there’s not one yet with regional skill, to investigate a highly localized phenomenon like UHI. Things that make you say WTF.

    Here’s a non-paywall version btw.

  73. Carrick said

    @Grzegorz:

    Guess what looks more convincing

    There’s “convincing” and there’s reality, which dictates what is, rather than what people want it to be.

    My discussion of tmin above simply demonstrates that confirmation bias exists within this community, and this bias is driving unwarranted conclusions about data quality.

    And these are comments from a guy who says “UHI can amount to 2°C total. Microsite effects can be much larger.”

    You can read what people say without drinking the koolaid they pour for their guests.

  74. @Carrick

    But you’ve noticed that the paper is an answer to McKitrick & Michaels (2007)? Because you missed the link to the full text under the abstract.

  75. @Carrick

    There’s “convincing” and there’s reality

    Damn right. The only question is why somebody’s “I have my doubts” should be closer to reality than decades of published research results. Because as far as “convincing” is concerned, the research results win hands down.

  76. Andrejs Vanags said

    Grzegorz Staniak provided a link (thanks) to:
    http://berkeleyearth.org/Resources/Berkeley_Earth_UHI.pdf

    I just started reading, but something caught my eye (section 4. Berkeley Earth Surface Temperature Global Average:
    “Stations are weighted according to their spatial distribution, taking into account their spatial correlation, so that
    regions with a high density of stations are not over weighted.”
    I dont understand that. Does that mean that regions that contain more data (and thus less uncertainty) are underweighted to the same level as the large sparse areas that contain less information?
    This is backwards. The sparse areas with a lower density of stations should be the ones underweighted.

    The other think I dont understand is the very large dropoff of temperature sensing stations worldwide, and the lack of calibration, maintenance and repair of the same. With the billions spent in analyzing the data form the stations, I believe more money should be placed on the stations themselves. The above paper mentions a peak of 39,000 sites (which by now are subtantially less).

    Another paper describing the “Berkely Earth Averaging Process” mentions that it uses the GHCN data set of 7280 weather stations but it mentions:
    “The sudden drop in the number of stations ca. 1990 is largely a result of the methodology
    used in compiling the GHCN dataset; GHCN generally only accepts records for stations that
    explicitly issue a monthly summary report however many stations have stopped reporting
    monthly results and only reported daily ones. Despite this drop, Figure 4(c) shows that the
    coverage of the Earth’s land surface remained above 95%, reflecting the broad distribution of the
    stations that did remain.”

    That does not make any sense, ignoring data because “there is no monthly summary” when the daily data exists? That should be corrected and the GHCN data base updated with the data from all those stations that it “dropped”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
Follow

Get every new post delivered to your Inbox.

Join 133 other followers

%d bloggers like this: