The Smoking Gun At Darwin Zero?

This is a guest post by AnOilMan


No, but it sure is an enticing title.

It was 4 years ago when I first started getting concerned about what climate change skeptics were saying. This was the first article foisted on me: The smoking gun at Darwin zero. It’s an article by Willis Eschenbach which claims that adjustments made to temperature records – also called homogenisation – is evidence that scientists are fudging the results to further a political agenda.

At the time, I reasoned that even if the data was fudged badly, it couldn’t affect the overall results because there are thousands of measurement stations and Willis’ article refers to just one station in Darwin. I felt that it claimed there was something nefarious going on without actually proving it. That’s when I decided that WattsUpWithThat was a waste of time.

Here’s the data at Darwin Airport if you’re curious. Here’s what they say about their data and here is the Paper outlining why this data was adjusted. The intro says:

“Consequently, each station record was corrected for discontinuities caused by changes in site location and exposure, and other known data problems (Peterson et al. 1998). Such discontinuities can be as large, or larger than, real temperature changes (e.g. the approximate 1°C drop in maximum temperatures associated with the switch to Stevenson screen exposure (Nicholls et al. 1996)) and consequently confound the true long term trend.”

It was the 1940’s when they installed the Stevenson Screens for weather stations, and if you look at Willis’ data, you can see a 1C discontinuity quite clearly. Hardly something to say Yikes over …

On another note, some years later, the fellow who showed me Darwin Ground Zero told me that he thought it was all a conspiracy to fake the data, just like UFO landings. He went on to tell me that only special people like him could spot it. I laughed …

So what does Temperature Data contain?

In a word, ‘errors’. A good temperature dataset is one which includes changes in climate only but sometimes there will be variations caused by non-climatic factors like, in the case of Darwin, the installation of Stevenson screen shelters for housing measurement instruments. Historically weather stations may also have had their IDs reused from other stations, they may have been moved, or their measurement apparatus upgraded. There are not always logs of adjustments to locations, so the data will just appear different at those locations. Thus in order to detect long-term trends in temperature datasets, non-climatic influences on temperature must be removed.

Victor Venema discusses this over at Variable Variability.

This article shows an example of why homogenization needs to be done. It’s not a great example in that this is probably from a different location, but you get the point. It shows what appears to be a cold spot in the middle of the dataset as though there is a sudden local refrigeration event like so:

How can they make a determination that the data is indeed wrong, and in need of repair? It’s simpler than it sounds, but as data and grid quality improve, it’s possible to see just how inhomogenous that location is. For instance, precipitation measurement used to be done on roofs of buildings until someone noticed that it reported precipitation levels that were 10% lower than measurements taken on the ground. If you move the station to the ground from the roof, there’s a jump in the data set.

Records of changes in measurement stations really help in making these determinations as is the case for ‘Ground Zero Darwin’ … 🙂

What is done to clean the data?

Homogenization is adjusting temperatures based on what other stations are measuring. Two stations nearby ought to show similar climate trends. If these trends diverge from each other, then this probably indicates the existence of non-climatic influences. The interesting part is figuring out which station has been affected by this influence.

It’s relatively easy to determine how accurate certain stations are relative to other stations using modern data which is known to be more accurate and has denser coverage. We can then go back in time using old data to reevaluate the measurements. The point is that you know approximately where a measurement is supposed to be, and thus you can identify outliers.

Victor Venema has a how-to on Statistical homogenisation for dummies. Of course, being a climate scientist himself, he may be biased. 🙂

Is this clear evidence of Willis’ assertion: “clumsy fingerprints of someone messing with the data Egyptian style … they are indisputable evidence that the “homogenized” data has been changed to fit someone’s preconceptions about whether the earth is warming.” Hardly, that statement is supposition.

But hey, what does this do to the global temperature measurement?

Not much. Here’s a great article at Skeptical Science comparing raw data to adjusted data. You can see a side-by-side comparison of global adjusted data with global unadjusted data in Figure 5:

Interesting … is the homogeneity adjustment artificially creating the ‘pause’ by amplifying El Nino? Can anyone answer that?

BEST uses automated statistical methods to arrive at similar conclusions.
Here’s a good break-down of computation methods used by BEST Berkeley Earth Temperature Averaging Process. When it comes to discontinuities in the data, they don’t homogenize, they simply break the time series, and evaluate the temperature sequences separately. Instead of one trend line, they will generate two, one for each time interval. (What is important is that they are trying to measure station temperature trends, not absolute temperature. Look up Scalpel in the BEST paper.)


AnOilMan is an electrical engineer who works in oil and gas.

 

Advertisements
This entry was posted in Climate change, Global warming, Watts Up With That, Wilis Eschenbach and tagged , , , , , . Bookmark the permalink.

59 Responses to The Smoking Gun At Darwin Zero?

  1. Rachel says:

    Thanks for posting this, AnOilMan. It’s a very interesting topic and a good example of how people can arrive at an incorrect conclusion when they learn just a little bit about the science. It’s like the saying, “a little knowledge is a dangerous thing”, which comes from Alexander Pope:

    A little learning is a dangerous thing;
    drink deep, or taste not the Pierian spring:
    there shallow draughts intoxicate the brain,
    and drinking largely sobers us again.

  2. Bobby says:

    Thanks for the post. I’ve read some of Victor’s articles, but hadn’t seen the “for dummies” one. Also, the pic that has the “local refrigeration event” says a lot.

  3. The pictures shown in this article may explain something even to you, who cannot read Finnish

    http://ilmastotieto.wordpress.com/2010/10/18/heikki-nevanlinna-ilmatieteen-laitoksen-historialliset-lampotilahavainnot-helsingissa/

    The last figure (Kuva 12.) shows raw (solid line) and adjusted data (dashed line) from weather stations of Helsinki. The location has moved a couple of times but all locations are inside a circle of 300 m diameter. (Earlier measurements from 1828-44 are not included, Their location is about 1 km south of these.)

    Some equipment of the present and previous equipment are shown in the figures 3-11.

  4. The vast majority of the adjustments over the past 100+ years have to do with calibrating SST records. Since SST accounts for ~70% of the global temperature series, this calibration is critical.
    Most skeptics don’t realize that the uncalibrated pre-WWII SST measurements ran significantly cold, and the calibration was to increase the temperature of those years after the corrections were applied

    The WWII years are the most uncertain as protocols changed abruptly.

  5. AnOilMan says:

    WebHubTelescope: It was post WWII when they started detailed temperature\salinity profile analysis because they realized that the seasons changed how easy it was to spot subs with sonar. They discovered that Temperature\Salinity guides sonar acoustics away from particular depths.

    This data is of course military, and the raw contains measurements for ships that ‘aren’t there’. Furthermore, to hunt subs, most allied ships and subs use the Levitus Database to determine the temperature\salinity profile for any region of the ocean. That information is homogenized by any recent XBT data. (Its impossible for a sub to generate its own XBT data, so it must use the accuracy of the database to hide.)

    Levitus wrote the book on extracting accurate temperature information from sparse geo-spacial and sparse temporal databases. If he’s wrong then the American navy sucks.

    But some people just want to bash cold war heroes;
    http://wattsupwiththat.com/tag/levitus/

    Pekka, Thanks for that post.

    I actually swapped out a thermometer at a military base. I replaced a 30 foot column of mercury that went from a proper (POST WWII) measurement station in the shade out of the wind, to the warm window of the MET Office. The replacement thermometer was digital, and placed in the sun at the end of the jetty, which communicated using an isolated 4-20 mA current loop. Before we cut over to the new thermometer the ranking meteorologist, looked at both measurements, added a linear offset so they looked similar.

  6. AnOilMan says:

    I think pseudo-skeptics focus on temperature records because they don’t understand them.

    But what exactly is the ‘smoking gun’? And who is it pointed at? In retrospect, Willis Eschenbach’s article looks pretty silly. Its full of frothing at the mouth political attack language, and weak on facts.

    Post Climate Gate.. fully exonerated the scientists;
    http://en.wikipedia.org/wiki/Climatic_Research_Unit_email_controversy

    “indisputable evidence that the “homogenized” data has been changed to fit someone’s preconceptions about whether the earth is warming”
    Yet BEST doesn’t manually adjust the data at all and still arrives at the same conclusion.

    I think that it speaks volumes about WUWT, and I don’t think its good.

  7. AoM,
    Interesting post, thanks. I haven’t covered this type of thing at all, so it’s very useful to have something like this included.

  8. Carrick says:

    Nick Stokes’s latest post:

    http://moyhu.blogspot.com/2014/04/march-giss-temp-up-by-025.html

    Has an inadvertent sensitivity test in it.

    Looking at all of the factors that influence the change in temperatures over time is pretty interesting. Nick and I briefly discussed some of the issues associated with that. Some of these are changes in the data sets, late reporting of data, changes in individual values due to quality control, and of course algorithm changes over time.

    Most people agree that temperature adjustments are important. They are crucial in fact to getting the correct answer. Most of the trend in the US temperature series comes from the adjustments for example. So making the adjustments, and getting the adjustments correct, is an important step in accurately assessing the change in global mean temperature over time.

    In fact, when Jeff Condon and RomanM developed their own index, correcting for a bias produced by the way anomalies are computed in most standard series, they found that the corrected algorithm produced a slightly higher trend. I believe Nick’s TempLS and BEST both use this improved algorithm.

  9. That is a pretty amazing Willis Eschenbach post. Well amazing if you are not used to the posts at WUWT and Co yet.
    They give so much space to showing Australian temperature observations that started before 1900, but somehow the space is missing to mention that Stevenson screens were introduced in Australia after 1900, if I recall correctly. Before that measurements were made on stands, that needed to be rotated to avoid the sun shining on the instruments. No wonder the observations in the 19th century were too high.
    And there is no attempt to try to understand the “findings”. Very unscientific, but very suggestive and good PR.

  10. Joshua says:

    “Most people agree that temperature adjustments are important. “

    Indeed. I read all the time, at WUWT, how important “most” people consider temperature adjustments to be.

    I love how some “skeptics” display their skepticism.

  11. Nick Stokes says:

    For me too, this post of Willis’ got my blog started. It’s hard to imagine how such a routine cherrypick got so much attention, though it was at the start of the Copenhagen conference. Anyway, ordinary political blogs like Kevin Drum, Volokh and Megan McArdle wrote posts. The Economist published a takedown, and some to and fro resulted.

    Darwin was an outlier – I pointed out at WUWT that Coonabarabran in NSW showed a similar but opposite response to adjustments. No-one was interested.

  12. It really is the difference between a random set of corrections, more-or-less equally weighted up-and-down versus a systematic bias in one direction over a long time span.

    That is why the SST corrections are so important. Look at slide 6 on this set:
    “Uncertainties in SST and Sea Ice Analyses”
    http://jra.kishou.go.jp/JRA-25/3rac/program/G4-442.pdf
    Before WWII the bucket correction was about 0.4C. This calibration is based on marine air temperature readings which were more sparse but didn’t suffer from a systematic bias.

    I don’t know if the temperature calibration during the WWII years will ever get straightened out. When I do fits to historical temperature rise using natural variability models, that is the time interval that is always the most problematic.

  13. BBD says:

    I think pseudo-skeptics focus on temperature records because they don’t understand them.

    They do it because they have invented an imaginary conspiracy amongst scientists to exaggerate and “promote” “CAGW” for the purpose of opening the doors to world socialism and the “redistribution of wealth” from developed to developing economies.

    Sure, they don’t understand the temperature reconstructions, but they don’t understand anything much about physical climatology either. They don’t really need to because their intent is to deny it not actually get to grips with how it works and what it means.

  14. Carrick says:

    Nick Stokes:

    For me too, this post of Willis’ got my blog started. It’s hard to imagine how such a routine cherrypick got so much attention, though it was at the start of the Copenhagen conference.

    I’m always a bit surprised to see how influential the usually wrong posts by Willard can be.

    Actually though, it seemed to me at the time that the release of the GISTEMP code and CRUTEM code & data drove a lot of blogosphere activity, in terms of new temperature constructions.

  15. Carrick says:

    By the way, Steven Mosher and Zeke Hausfather also discussed temperature adjustments on WUWT. It also includes a partial list of temperature reconstructions. There are some decent comments there too.

  16. AnOilMan says:

    Carrick, Nick Stokes: I seem to recall reading an article on SKS about how temperature trends are calculated. I can’t find it. It was good because it graphically showed how they arrived at the methodology being used today, an average trend line calculated by location. Is there any chance you know where that is, and what I’m talking about?

    BBD: I don’t think that they all fit the ‘conspiracy theory’ explanation. I don’t think you need to be a conspiracy theorist to fall for Willis’ sucker punch. Most people haven’t got a real understanding of physics, let alone graduate level physics, intermixed with geo-spacial statistics, and a thorough understanding of meteorology.

    I really did learn a lot in reading up on this.

    What usually ticks me off is the inflated political language. That more than anything else shows an ulterior motive. Language is almost always what I look at first. If Willis had said, “Hey, what’s this?” that would have been more appropriate.

  17. BBD says:

    AOM

    I seem to recall reading an article on SKS about how temperature trends are calculated. I can’t find it. It was good because it graphically showed how they arrived at the methodology being used today, an average trend line calculated by location.

    Have you see, this?

    * * *

    The longer I look at pseudosceptics, the more I see ‘conspiracist ideation’ and projection as being near-universal traits.

  18. BBD says:

    Willis’ post reeks of suspicion:

    What this does show is that there is at least one temperature station where the trend has been artificially increased to give a false warming where the raw data shows cooling. In addition, the average raw data for Northern Australia is quite different from the adjusted, so there must be a number of … mmm … let me say “interesting” adjustments in Northern Australia other than just Darwin.

    And with the Latin saying “Falsus in unum, falsus in omis” (false in one, false in all) as our guide, until all of the station “adjustments” are examined, adjustments of CRU, GHCN, and GISS alike, we can’t trust anyone using homogenized numbers.

    Regards to all, keep fighting the good fight,

    w.

    He’s a conspiracy theorist.

  19. The longer I look at pseudosceptics, the more I see ‘conspiracist ideation’ and projection as being near-universal traits.

    As much as I agree with AoM that not all who buy the pseudo-skeptic arguments are consciously being conspiracy theorists, it is hard to see how the pseudo-skeptic arguments can be right unless there is a major conspiracy. Hence, there is certainly some kind of default position where those who buy into the pseudo-skeptic arguments are effectively buying into a conspiracy theory, even if they don’t quite realise it.

  20. Carrick says:

    AnOilMan:

    I seem to recall reading an article on SKS about how temperature trends are calculated.

    I’m sure what you’re looking for. Are you talking about the estimation of the trend itself?

    Most people haven’t got a real understanding of physics, let alone graduate level physics, intermixed with geo-spacial statistics, and a thorough understanding of meteorology.

    That’s definitely the case. Same goes for actually going through the adjustment (homogenization) process so you understand how the machinery works.

    The thing is, it’s eventually the algorithms and what they are actually doing that matters, not the verbiage that accompanies it. And there is usually a good amount of daylight between the procedures actually used and the documentation that accompanies them, let alone narratives based on out-of-date documentation that further distort what is actually happening.

    As a for instance, Just because I wrote the code doesn’t mean I understand all of the nuances about how it functions in practice, when my algorithms meet data errors I didn’t plan for let alone fully analyze the consequences of. The reality is what the code does not what I say or think it does.

    If Willis had said, “Hey, what’s this?” that would have been more appropriate.

    I agree… more humility would be appropriate.

  21. Carrick says:

    Sorry that should read “I’m not sure what you’re looking for. Are you talking about the estimation of the trend itself?”

  22. BBD says:

    AOM

    Re: the SkS post on how trends are calculated, was it this?

  23. AnOilMan says:

    “Stay tuned for next week when middle aged man, can’t remember something or other.” Nope those aren’t the articles. It was how they calculated the average trend line, and how they arrived at the final calculation methodology, and what it compensates for like sparse or irregular data collection. It may not have been SKS.

    BBD: I’d bet Willis’ article is intended to appeal to a conspiracy theorist, but it could and did fool a lot of people.

    During the last Canadian elections a kid posted in another forum some article filled with hate derision and anger, and asked if it was true. It was a tour de force barrage on Stephen Harper. And while I’m no fan of Harper, me and several others (all Liberals actually) took the time to explain what that material was and that it was no basis for making a decision.

    That kid was in high school and lacked the critical thinking skills necessary to evaluate what he saw on the internet.

  24. Carrick says:

    AnOilMan:

    Nope those aren’t the articles. It was how they calculated the average trend line, and how they arrived at the final calculation methodology, and what it compensates for like sparse or irregular data collection. It may not have been SKS.

    Well if you figure it out, please post the link here, if you figure it out.

    Hu McCulloch has a description here on correcting for serial correlation.

    Nick has a post here that leverages on Hu’s post.

    I’ve got an analysis like that, but I’ve never written it up. However, I think I can explain how to do it well enough, even if I can’t use embedded equations here:

    Basically, you compute the trend by latitude band (“zonal trend”) as was done here. (This is technically for the range Jan 1979-Dec 2011 inclusive btw.)

    Then you compute the global trend by computing the cos(latitude)-weighted average of the zonal trend. That is compute the sum of trend(latitude) * cos(latitude * pi/180) over latitude. Then divide this by the sum of cos(latitude * pi/180) over latitude.

    If you do this for a series that robustly estimates the temperature for regions with missing data, e.g., Cowtan & Way (2013), you get virtually the same result:

    From their published global series, I get the trend for Jan 1979-Dec 2011 to be 0.18091 C/decade. Using the weighted average of the zonal trends, also gives 0.18090 °C/decade. (I’ve retained the results to the first digit where they are different.)

    On the other hand, using the global average published by HadCRUT gives 0.149°C/decade.

    The weighted average of the zonal trend however gives 0.165°C/decade.

    GISTEMP for the same period gives 0.209°C/decade and NCDC gives 0.156°C/decade.

    Hope this helps!

  25. Carrick says:

    These:

    GISTEMP for the same period gives 0.209°C/decade and NCDC gives 0.156°C/decade.

    Are the weighted averages of the zonal trends.

    The trends based on their published global series are respectively 0.167°C/decade (GISTEMP) and 0.155°C/decade (NCDC).

  26. AnOilMan says:

    Found it…

    This series of articles clearly shows how the trend calculations are made, why they are done that way, and what they are compensating for.
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_1A.html

    I like this article because it is ‘accessible’ to those who are less technical. And it shows how the measurements arrived at the techniques being used today. All too often people think that scientists just invent formulas and use them. They don’t get the whole picture that it took decades of getting it wrong before finding the right answer.

    After you stare long enough your eyes get sore, and you notice a warning label that came with your monitor… “Remember to blink.” The wife is mad.. .something about missing kitchen sheers, and we need propane.

  27. Eli Rabett says:

    The cleverest way of figuring out how good homogenization is, is Tom Karl’s Climate Reference Network, described at the link below with some broken links, but use your imagination. Atmoz was a good blogger.

    http://rabett.blogspot.com/2010/01/toms-trick-and-experimental-design.html

  28. jsam says:

    Someone has bitten the ears off the bunny’s second and third images, Eli.

  29. Carrick says:

    Not peer reviewed, but still interesting.

    Earth Atmospheric Land Surface Temperature and Station Quality in the Contiguous United States by Muller et al (2012).

    This was a contribution for The Third Santa Fe Conference on Global and Regional Climate Change.

  30. Steve Bloom says:

    It seems to have passed through peer review.

    But that reminds me: Has anything been heard of the contemporaneous Watts-authored blockbuster, or has its residence in the memory hole become permanent? IIRC it noisily failed on the boneheaded amateur error of failing to account for TOBS changes, but even more fundamentally it sought to do something with Leroy’s work that was entirely unsupportable, i.e. it assumed maximum potential station errors to be the same as actual errors. But what else could we expect, that being the founding mistake of the entire surface stations project?

  31. AnOilMan says:

    Most of their work goes dark, or falls back behind a pay-wall when its proven utterly wrong. But it remains here;
    http://www.populartechnology.net/2009/10/peer-reviewed-papers-supporting.html
    (Keeps up the stats don’t you know.)

    Interestingly, other Watts papers aren’t there, like the one he did with Muller disproving UHI. Watts Up With That I wonder?

    Last time I went looking there, those papers were in journals with dog horoscope statistics, and analysis of likely UFO landing sites. (Mostly I’m saying many of these papers aren’t well reviewed. I don’t wish to offend any dog aficionados, or aliens.)

  32. Carrick says:

    AnOilMan, while I have issues with Fall 2011, what are the problems specifically you are concerned about?

  33. Carrick says:

    Fall 2011 is located here.

    Figure 4, panel 1 is the main result IMO. I think the other panels are just really looking at how urban vs rural affects the diurnal cycle. You need to look at the trend on the 24-hour averaged period (panel 1) to see if there is a net bias on trend.

    It’s actually pretty impressive how well the adjustments do, when you compare CRN 1&2 vs CRN 5. I would have expected a bigger effect.

    Steve Bloom—I could be wrong, but I think the journal just published the proceedings of that conference. I thought the paper had the feel of something that wasn’t fully prepared for regular peer review, in any case.

  34. AnOilMan says:

    Carrick… you give me to much credit. I tend to ignore people who just plain don’t know what they are doing. That would be Watts and crew. Reading anything from them is like a big wet willy in the ear, and not my favorite experience. As they dig out more technical explanations its just more work to read.

    Muller seems to be arguing Fall 2011 is wrong. And specifically saying that the difference aren’t statistically significant, by measuring it, and pointing out that Watts data uses a poor sampling of regions. i.e. cherry picking stations so their measurements over amplify the problem.

    I’d want to know what went into the selection for the Fall 2011 stations observed. I’ve seen enough shenanigans from them to wonder if the data was fudged.

    Barring that, I’d wait till some real professionals take a look at it. Watts’ high school degree can only carry this so far IMO.

  35. Carrick says:

    AnOilMan, I don’t think there’s anything seriously wrong with the data in Fall 2011 (which includes other people that a professional meteorologist, including several prominent climate scientists), and I don’t think Fall 2011 was wrong was really the point of Muller 2012 either: They go out of their way to not directly address the minimum/maximum temperature differences, and there’s nothing controversial about the CRN comparisons for daily mean temperature.

    Fall 2011 uses the surfacestations.org classification of USHCN stations (the data set developed by Watts), as does Muller 2012 and Menne 2010. This is the paper that Eli was mentioning in his article, with the broken-url hot-linked images (hot-link karma).

  36. Steve Bloom says:

    It has submitted and accepted dates, which I take as indicating peer review (FWIW).

  37. Pingback: Another Week of Climate Disruption News, April 20, 2014 – A Few Things Ill Considered

  38. dhogaza says:

    Steve Bloom: as a matter of fact, yes, the “Watts blockbuster” is being heard from once again.

    Search this thread for Evan Jones, you won’t be disappointed. Still looking for ways to justify throwing out station data that doesn’t fit their pre-conceived notion:

    http://blog.hotwhopper.com/2014/04/hotwhopper-competition-best-name-for.html

    Warning … Evan posts a lot.

  39. Carrick says:

    Steve Bloom:

    It has submitted and accepted dates, which I take as indicating peer review (FWIW).

    I see what you mean. I think you are right about it being peer reviewed (FWTW).

  40. Carrick says:

    I agree with Evans on this:

    I have simply dropped all TOBS-biased stations (the bias is real but I don’t trust the adjustments)

    If you want to look at individual station records rather than an aggregation of them.

    The method used for TOBS wasn’t designed to work for individual stations, but I think it works well enough for what it was intended for—aggregation of data.

    I saw John N-G is still working on the paper so that’s encouraging.

  41. AnOilMan says:

    Carrick: Pretend you’re educating me. This really isn’t my thing. What is your concerns with those papers?

    I’m not sure if BEST actually processes minimums and maximum temperatures, and more than anything else I bet he was concerned that his previous work was correct. Of course… if BEST showed the issues to be worse, then he may remove that bit…

  42. Steve Bloom says:

    dhogaza,

    “Nobody!
    “Nobody!
    “Nobody out-bloviates the Rev!”

    🙂

  43. Carrick says:

    BEST does tmin/tmax too. Now I just need them to do an hourly. 😉

    I think the data and methods in Fall 2011 is good (on a poor/fair/good/excellent scale). But I think the analysis is fair to poor.

    The biggest concern I have is the observer reports are snap-shots of the condition of the station at that point in time. Interviews were performed in some cases, but it does not appear to have been done systematically, nor properly documented.

    My problems with the analysis is that it is being done without a clear understanding the various adjustments are being made, and separately how the microclimate affects the daily temperature cycle, without necessarily producing a change in trend, for data that is average over an annual cycle.

    I think the following is true, and I think it illustrates the problems with the apparent lack of model-driven analysis on the part of Watts and company.

    Assumptions:

    • Assume you have a station with given microsite properties, but the microsite properties are stationary in time.

    • Also, assume that your site is exposed to normal wind circulation patterns (e.g., not down in a well), so it can be considered to be sampling atmospheric temperature.

    Then:

    I think on conservation of energy+equilibrium approximation grounds, you’ll find that there may be a bias in the temperature offset of the microsite compared to the region-scale annually averaged 1-m above surface temperature. But you won’t get a difference in trend over annually averaged temperature. More technically you have a non-zero offset bias, but zero scaling bias.

    In that situation, which is typical, the anomalization process exactly cancels out any micrositing issues. Hence, Fall Figure 4, left panel shouldn’t be a surprise.

    Since you don’t know the regionally averaged temperature, you don’t know the true value of the bias, but it doesn’t matter as long as the environment is stable at the measurement site.

    Changes in the the offset due to e.g. a new building being added doesn’t introduce a scaling bias, rather it produces a “step-function” in temperature change. But it won’t cause the temperature to continue to climb at one site compared to another. That’s just unphysical.

    Homogenization algorithms can detect and correct for these types of step-functions IMO, BEST has done very well in advancing the state of the art here. IMO, it’s completely bone-headed to not make homogenization adjustments, if what you want to measure is temperature trends on time and spatial scales relevant to climate change.

    So realizing that changes in micrositing can influence the offset bias, what you should really trying to do with your study is try and improve the homogenization, TOBS and other corrections.

    But you can’t do that if you don’t grapple with the hard questions of how these various effects influence your offset bias.

    And I think you are really in trouble if you don’t understand what is required to really observe different temperature trends. One shouldn’t get lost in the intricacies of the math on the one hand, but you should keep in mind what basic physics imposes on your system too.

    I’m already in tl;dr territory so I’ll stop here.

  44. Pingback: Another Week of Climate Disruption News, April 20, 2014 [A Few Things Ill Considered] | Gaia Gazette

  45. AnOilMan says:

    Carrick: Thanks for that.

    I had already assumed that was the issue with much of this focus on absolute temperature readings. Specifically we don’t need an absolute accurate reading if all we are looking for is a trend line.

    The simple math here helped me understand all that.
    http://www.skepticalscience.com/OfAveragesAndAnomalies_pt_1A.html

    And I did wonder how the sites were evaluated. Ad hoc observer reports have another name, ‘Qualitative’. Its not really a compliment.

  46. Eli Rabett says:

    jsam Atmoz is vanished and his blog with him;)

  47. Steve Bloom says:

    NCDC was originally going to do the CRN in pairs, but when they did a preparatory experiment they discovered that identical instruments placed a very short distance apart with identical site conditions nonetheless gave noticeably different readings (although of course the anomalies were fine). RP Sr. and Willard Tony were uninterested in this information, nor did they seem to want to just wait for enough CRN data to use to validate (or not) the USHCN data (noting the CRN/USHCN station pairing designed to facilitate exactly that). Their entire effort never had a thing to do with science. The intentional misuse of Leroy’s classification system was just the icing on the cake.

    BTW, from Leroy (2010): “Unfortunately, the siting classification as it is defined, doesn’t allow to
    correct the measurements.” Indeed.

  48. Steve Bloom says:

    Oh, and the following sentence: “Correction methods remain possible, but independently of the siting classification.”

  49. Carrick says:

    AnOilMan, the site evaluation used by surfacestations.org is from the NCDC guidelines.

    I don’t think qualitative is necessarily a problem, as long as the selection guidelines are good enough.

  50. Eli Rabett says:

    Selection makes close to no difference at all in anomalies, which is one of the lessons of BEST, but was well known before that.

  51. Carrick says:

    Eli:

    Selection makes close to no difference at all in anomalies, which is one of the lessons of BEST, but was well known before that.

    With correct adjustments, yes.

    See Menne 2009 & 2010 for example. This is also (IMO) the main result of Fall 2011. Sorry for the severe butt-hurt, Anthony. 😉

  52. AnOilMan says:

    Stoat has a more entertaining way of summing BEST up;
    http://scienceblogs.com/stoat/2012/07/28/muller-is-still-rubbish/

    “no, your actual findings are simply the same as IPCC 2007: all the UHI stuff, and the data selection issues: its been done before.”

    Or better yet, Mann;
    “At this rate, Muller should be caught up to the current state of climate science within a matter of just a few years!”

    I keep worrying that Muller will do an about face like so;
    http://berc.berkeley.edu/dr-richard-muller-author-of-energy-for-future-presidents-speaks-on-evs-and-natural-gas/

  53. Steve Bloom says:

    TBC, NCDC adopted the Leroy scheme. TB even more C, it was for siting the CRN, not for doing quality control on the HCN.

    I suspect AOM’s remark about qualitative had to do with the likes of Evan Jones making the determinations based on data collected by volunteers.

  54. Carrick says:

    AnOilMan, regardless of snotty comments by crusty old scientists, things are what they are. Three examples:

    • The kriging algorithm used by BEST is a substantial improvement over what NCDC, HadCRUT or GISTEMP have done.

    • The homogenization algorithms of BEST is a substantive improvement too.

    • And of course, no other series offers daily homogenized temperature series besides BEST, that includes Tavg, Tmin and Tmax.

    Steve Bloom:

    I suspect AOM’s remark about qualitative had to do with the likes of Evan Jones making the determinations based on data collected by volunteers.

    The surfacestations.org data is actually quantitative that was collected… documentary photographs, distance measurements, etc.

    (Technically CRN1-5 is categorical data, not qualitative data, by the way.)

  55. Steve Bloom says:

    Sure, but I suspect not enough to do the categorization correctly, which makes the latter questionable as any sort of data.

  56. Carrick says:

    It doesn’t seem to be that difficult to accurate determine which category a station falls into.

    Getting a gps reading of the site allows GoogleEarth to be used to assay the location. (The recorded values for most stations were not accurate enough to reliably identify the station, especially when the station was mixed in among buildings.)

    Taking a series of photos from different angles provides up close details, and to verify the location of the station relative to nearby buildings, and to identify potential heat sources.

    It appears that distances were measured using GoogleEarth, using the ground truth obtained from the observers.

    Menne and others have used versions of this data set, so it does appear that they were satisfied with the quality of the station survey.

  57. “But that reminds me: Has anything been heard of the contemporaneous Watts-authored blockbuster, or has its residence in the memory hole become permanent?”

    It is currently discussed at HotWhopper.

  58. Carrick are you part of the BEST team?

  59. Carrick says:

    Victor, no, I’m not part of BEST.

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