Petaluma CA

Check out Petaluma CA at Anthony Watts’ blog.

Anthony has shown the GISS unadjusted temperature graphic. In this case, the GISS raw data appears to be mostly the same as USHCN adjusted (filnet) up to different rounding; a few isolated values available in USHCN are missing in GISS raw for some strange reason. I’m starting to keep track of the GISS raw source, as I pick up these files. I’m going to check whether the USHCN filnet is consistently picked up or not. At any rate, it is here.

The GHCN raw version is the same as USHCN raw version and differs from the GISS raw version. The GHCN adjustment has the effect of lowering early 20th century temperatures relative to modern temperatures. The GISS adjustment goes the opposite direction from the GHCN adjustment. Two figures are shown below. First the raw USHCN minus the three adjusted versions: USHCN, GHCN and GISS and then the amount of the GISS adjustment relative to the GHCN adjustment, which reaches up to 1.5 deg earlier in the century.

In the cases that I’ve examined, Parker uses GSN data (which appears to be equivalent to GHCN raw data where they overlap.) In this case, Hansen adjusts for 1.5 deg of urban warming that Parker has demonstrated not to exist. If Parker is right then the GISS data is seriously understating 20th century warming by adjusting for a supposed urbanization trend effect that Parker has “demonstrated” is not present in the data.

Figure 1. Adjustments to Petaluma (in deg C)

Here are the 7 versions overlaid:

Figure 2. Petaluma CA versions

The station history at CDIAC is not up to date showing moves with directions are shown in 1927, 1941, 1954, 1968, 1969, 1981 and 1991, but not two recent moves listed at MI3. Obstructions at the present site are listed at MI3 as: “MMTS 259/35 F&P 055/10 TREES 001-030/30-45/15-3 TREES 150-210/320-180/3-5 TREES 240-330-359/40-30-25/12-13-15 BLDG 280-330/45-55/6-6ll”. Go to Anthony’s pictures for more details on the obstructions described on the form so laconically.


  1. Steven mosher
    Posted Jun 20, 2007 at 12:07 PM | Permalink

    I’m Apopleptic. that’s like gobsmacked squared.

  2. Posted Jun 20, 2007 at 12:22 PM | Permalink

    This is probably an obvious question, but do we know what methods and criteria were used to create any of these adjustments? Ad hoc secret?

    Hansen claims that his curve fitting models show that only CO2 increases can explain the surface temperature history of the 20th century. The surface temperature history seems rather fluid (and unaudited).

  3. Steve McIntyre
    Posted Jun 20, 2007 at 12:29 PM | Permalink

    #2. There is a little and very poor methodological description in some articles, but my impression is that the methods would be difficult to emulate based on these descriptions. I’ve requested source code from GISS and got no response. I requested source code from Vose at NOAA and got a very polite answer (as all of Vose’s emails have been) but he said that I’d have to try someone else (it will probably end up being Karl, who I expect to be uncooperative.)

  4. Byron
    Posted Jun 20, 2007 at 12:36 PM | Permalink

    I suppose we can be grateful that the adjustments have not been uniformly reconciled and we can still see what a pigs breakfast the whole process is.

    btw. great snapshot

  5. Steve Sadlov
    Posted Jun 20, 2007 at 12:43 PM | Permalink

    Until the mid 1970s Petaluma was a rural service center with a slow growth rate. Based on the norms of times previous to this, it was considered too far from San Francisco and other urban commercial nodes to be considered a viable commuter suburb. Back then, the suburban hot spots were either on the contiguous landmass to the south of SF, across the Bay to the East, or the low lands of (in)famous Marin County. By the mid 70s, with the exception of Novato, the Marin lowlands were all developed as were most other lowlands to the south and immediate east of SF. Furthermore, Marin now had a significant pastiche and hence, became innately inflated in land value. Leapfrog development ensued. Petaluma, being just over the border into Sonoma County, was the logical first destiation for the leapfrog developers. Today, there is a major automall, two regional shopping malls and myriad cookie cutter suburban tracts.

  6. pochas
    Posted Jun 20, 2007 at 2:11 PM | Permalink

    Steve M:

    In the cases that I’ve examined, Parker uses GSN data (which appears to be equivalent to GHCN raw data where they overlap.) In this case, Hansen adjusts for 1.5 deg of urban warming that Parker has demonstrated not to exist.

    I think this is a case of which came first, the chicken or the egg. Let’s assume that UHI is a local effect that appears full-blown when the urban environment is created. If a temperature sensor is then placed in the urban environment it will show no UHI trend, since the UHI effect was fully developed at the time the sensor was placed. I believe this is almost always the case and may explain Parkers findings. But the station readings nevertheless contain a UHI component which contributes when readings from a large number of stations are averaged.

    Still, it is not appropriate to adjust station readings to remove an effect which GISS is assuming to be variable but is actually practically constant, as I believe this is what Parkers paper actually shows.

  7. Steve Sadlov
    Posted Jun 20, 2007 at 2:41 PM | Permalink

    Imagine a very urban setting such as Manhattan.

    In 1950, there were shared telephones, no TVs, very few electrical gadgets, and much less electrical current in transit around town.

    Now, there are land and cell phones, a TV in every room, myriad gadgets, much more current and many more wires and cables carrying it. Older 3 and 4 storey buildings replaced with 40 storey glass towers. Jumbotron type TVs in Times Square. Etc.

  8. Posted Jun 20, 2007 at 3:16 PM | Permalink

    I would like to see dates for adjustments. Maybe this isn’t available but if the temperature reading of the thirties were adjusted down in the thirties it means something very different from adjustments made in the nineties. Making the thirties cooler in the nineties suggests of an agenda.

  9. Steve McIntyre
    Posted Jun 20, 2007 at 3:25 PM | Permalink

    All the adjustments are 90s or later.

  10. aurbo
    Posted Jun 20, 2007 at 5:19 PM | Permalink

    Re #7;

    I grew up in NYC in the 30s and 40s and during WWII when weather information was classified (!) I would haunt the guys at the City Office in order to satisfy my curiosity and interest in weather.

    In 1958 the (then) Weather Bureau published as part of their Climatography of the United States, a 61-page publication (Number 40-26) called “Climatic Guide to New York and Nearby Areas.” Although long out of print I still have the copy I obtained nearly 50 years ago.

    NYC weather observations were generated on a regular basis since 1821. The two main sites in NYC during the period up until the 1958 date of this publication were the City Office, located originally at Fort Columbus Marine Hospital in NYC was moved to Lower Manhattan at the inception of the US Weather service ion 1870. From its initial location at 49 Wall Street, and some nearby sites, it found its way about ⻠mile to the Southwest in the Whitehall Building at 17 Battery Place where it remained into the 60s.
    The Weather Office was on the 29th floor, but the instruments were on the roof a couple of floors above.
    Ground Level was 10ft ASL. The instruments on the roof ranged from 398ft AGL (rain-gauge) to 454ft AGL.. The temperature measurements came from instruments inside a thermoscreen (essential a CRS type shelter) 414ft AGL. The building was bordered almost three sides by the waters of NY Bay and the Hudson River.. It would be hard to find a less representative site for recording temperatures for NYC.

    The other site in Manhattan was located in or adjacent to Central Park. It originally bordered 5th Avenue with the instruments on the roof of the arsenal between 63rd-64th St from 1868 until 1920, at which time it was moved to Belvedere Tower well inside the park, but close to the 79th/81st St Transverse Road. The thermoscreen was 6ft above ground level (138ft ASL) on a well manicured grassy area about 20-30ft from the building. The book describes this location as follows: “This site is the highest ground in the park, and gives a generally excellent exposure for the instruments with the advantage of a more rural environment.”
    The site is still there, but operating unattended with readings transmitted remotely by wire to wherever the NYC WFO happens to be at the time.

    The following notation accompanies both temperature records: “Long term means were computed and the indicated extremes selected without regard to changes in location during the period of record. Before utilizing this data for exact analysis, the reader should consider indicated location changes carefully.”

    How come they had more respect for climate sensitivity 50 years ago than what appears to be the case now?

  11. Steve Sadlov
    Posted Jun 20, 2007 at 5:46 PM | Permalink

    Did the Ukiah thread get pulled? I don’t see it anywhere on the blog.

  12. Sam
    Posted Jun 20, 2007 at 5:46 PM | Permalink

    Because they’re all a bunch of idiots on a crusade

  13. jae
    Posted Jun 20, 2007 at 5:55 PM | Permalink

    10: See here. It shows a definite increase in dT/dt, compared to rural areas, even in Central Park.

  14. JeffB
    Posted Jun 20, 2007 at 6:24 PM | Permalink

    It would seem to me that the UHI effect would be a rather gradual phenomenon. It’s not like development just pops up over night! I would be suprised if a UHI effect could actually be teased out of a temperature record at a constant adjustement. IMO, it would in itself constitute a modelling effort in itself.

  15. Posted Jun 20, 2007 at 6:24 PM | Permalink

    Steve Mosher wrote:

    “I’m Apopleptic. that’s like gobsmacked squared.”

    If this goes on, you’re going to have to invent a whole new vocabulary to express your level of

  16. Anthony Watts
    Posted Jun 20, 2007 at 6:29 PM | Permalink

    FYI the website is back up and running!

    There is a response delay in the image database server, which we are working to resolve. But you can use the entire website now without my worrying about our office network getting smkacked by traffic.

    I expect we’ll figure out what the latency delay for the database server is tomorrow.

  17. Steven mosher
    Posted Jun 20, 2007 at 7:29 PM | Permalink

    RE 15.

    Like most things in life, outrage follows a logistic curve.
    So, Apopleptic ( spell check Isle 3, spell check isle 3!) is
    pretty much near the second knee of the curve and from here on out
    my meter is pegged. I’m a gnat’s ass from being thresholded.

    Seriously big citylib, all the AWG stuff aside, nobody of good conscience can
    can look at these sites and be very comfortable with the results that are derived from them.

    Climate science does not need its Eoanthropus dawsoni at this time. Just thought I’d toss you
    a bone.

  18. Posted Jun 20, 2007 at 7:36 PM | Permalink

    17#: Peterson 06 did not seem to worry too much about these sites

  19. Anthony Watts
    Posted Jun 21, 2007 at 8:24 AM | Permalink

    RE18 That’s exactly why its tremendously important to survey them all.

    By not checking the point of data collection and “assuming” that the weather station meets the published NOAA and WMO standards appears to have been standard practice for many researchers.

    If you were conducting an experiment where the results were likely to shape national and world policy, wouldn’t it be prudent to check the origin of the data set?

    Government (NCDC, Karl, et al) was charged with providing a relatively homogenous data set. Yet like many govenment programs, it fell off the radar of due diligence. The problem is real; some of the measuring stations (perhaps many) have not been properly kept up nor properly quality control checked. It’s a typical failure of bureaucracy that we’ve seen in many levels of government agencies.

    If accurate science is to be done, accurate data is needed. That is a fundamental part of the scientific process.

  20. John Nicklin
    Posted Jun 21, 2007 at 10:29 AM | Permalink

    Why worry about the quality of the sites, they show what Hansen et al want so they must be good. The ones that don’t comply can be subjected to data torturing until they submit and produce the proper curves. If no amount of manipultion will do the job, just ignore the offending site.

    The whole thing stinks. With the amount of bad data coming in from compromised sites, I’m surprised that anyone can believe the results that GISS, HadCRU and others put out. The problem is that nobody knows how tortured the data is; outside of a very few people, mostly CA readers.

    Gobsmacked and Apoplectic may be to mild to describe the situation. We may need to invent a new superlative for shock and awe.

  21. pochas
    Posted Jun 21, 2007 at 11:02 AM | Permalink

    #7 Steve Sadlov,

    As re New York City, I just ran across this on the NWS Central Park station. Note that that once again it is the UHI “adjustment” that generates the hockey stick and not the raw data, which shows some warming early in the century and none thereafter. I think this supports Dr. Parker.

    Also, I think I can imagine the PDO wobble in the raw smoothed January data – on the east coast yet!

  22. pochas
    Posted Jun 21, 2007 at 11:05 AM | Permalink

    Re #21 – Oops – forgot the link:

  23. pochas
    Posted Jun 21, 2007 at 11:10 AM | Permalink

    Here is the link:


  24. Steven mosher
    Posted Jun 21, 2007 at 11:11 AM | Permalink

    re 19 & 20.

    You can find a link to Peterson 2006 over on Peilke. Basically, Peilke/Davey wrote about
    some questionable sites in Colorado ( Anthony, your inspiration right?) So, Peterson
    Came back and analyzed the Peilke sites and concluded ” no microsite effects”
    The argument turns on Homogeniety adjustments.. Opaque to me at this point, I did a quick
    skim. SteveM the greater indicated that he may post something on it.

    Peterson was “somewhat” careful not to generalize conclusions.. But mumbled about weight
    of evidence building.

    However, toward the end Peterson says this.. Paraphrasing ” although I showed this doesnt matter,
    Thanks to Peilke and Davy. And This in no way excuses shoddy sites”

  25. Paul Linsay
    Posted Jun 21, 2007 at 12:34 PM | Permalink

    #21, 23, pochas,

    Is that adjustment in the second figure what I think it is? Is this added to the raw data? Are they really dropping the temperature by 4 degrees from 1909 to 1989 then increasing the temperature by 5 degrees from 1989 to 2005?

  26. pochas
    Posted Jun 21, 2007 at 12:55 PM | Permalink


    That’s the way I read it. Gobsmacks ya, doesn’t it?

  27. MarkW
    Posted Jun 21, 2007 at 2:20 PM | Permalink

    Amazing how large these temperature adjustments are. This site is not unique. Yet they claim to be able to find a signal, that is 1/15th the size of the adjustment that they are making.

  28. Earle Williams
    Posted Jun 21, 2007 at 3:09 PM | Permalink

    And that signal corresponds amazingly to the adjustment itself, compared to the apparently signal-free raw data.

  29. BarryW
    Posted Jul 22, 2007 at 9:06 PM | Permalink

    The USHCN site has a general description of the adjustment methodology.

    The homogenization algorithm is described as:

    1. First, a series of monthly temperature differences is formed between numerous pairs of station series in a region. The difference series are calculated between each target station series and a number (up to 40) of highly correlated series from nearby stations. In effect, a matrix of difference series is formed for a large fraction of all possible combinations of station series pairs in each localized region. The station pool for this pairwise comparison of series includes U.S. HCN stations as well as other U.S. Cooperative Observer Network stations.
    2. Tests for undocumented changepoints are then applied to each paired difference series. A hierarchy of changepoint models is used to distinguish whether the changepoint appears to be a change in mean with no trend (Alexandersson and Moberg, 1997), a change in mean within a general trend (Wang, 2003), or a change in mean coincident with a change in trend (Lund and Reeves, 2002) . Since all difference series are comprised of values from two series, a changepoint date in any one difference series is temporarily attributed to both station series used to calculate the differences. The result is a matrix of potential changepoint dates for each station series.
    3. The full matrix of changepoint dates is then “unconfounded” by identifying the series common to multiple paired-difference series that have the same changepoint date. Since each series is paired with a unique set of neighboring series, it is possible to determine whether more than one nearby series share the same changepoint date.
    4. The magnitude of each relative changepoint is calculated using the most appropriate two-phase regression model (e.g., a jump in mean with no trend in the series, a jump in mean within a general linear trend, etc.). This magnitude is used to estimate the “window of uncertainty” for each changepoint date since the most probable date of an undocumented changepoint is subject to some sampling uncertainty, the magnitude of which is a function of the size of the changepoint. Any cluster of undocumented changepoint dates that falls within overlapping windows of uncertainty is conflated to a single changepoint date according to
    1. a known change date as documented in the target station’s history archive (meaning the discontinuity does not appear to be undocumented), or
    2. the most common undocumented changepoint date within the uncertainty window (meaning the discontinuity appears to be truly undocumented)
    5. Finally, multiple pairwise estimates of relative step change magnitude are re-calculated at all documented and undocumented discontinuities attributed to the target series. The range of the pairwise estimates for each target step change is used to calculate confidence limits for the magnitude of the discontinuity. Adjustments are made to the target series using the estimates for each discontinuity.

    For UHIE they state:

    “the change-point detection algorithm effectively accounts for any “local” trend at any individual station. In other words, the impact of urbanization and other changes in land use is likely small in HCN version 2.”

    If I’m following this correctly they are looking for discontinuities relative to the nearby sites, but if all the sites are showing UHIE then how could this method adjust for it? The best you could get, I would think, was something that looked like the average slope of the line trend lines. Or is the assumption that UHIE appears only as discontinuities in the data? It would seem that if the majority of a data set is bad then it would contaminate the whole set, “correcting” the good sites based on bad data.

    They reference a number of papers for their algorithm but I can’t get to them since they are not available without paying for them. You would think they would have the algorithms online if not the papers. Although, actually I’d prefer to see the code.

  30. Steve McIntyre
    Posted Jul 24, 2007 at 10:37 AM | Permalink

    I agree with your concerns about the methodology. It’s a typical climate science recipe – a procedure that does not contain any references to known statistical methodology and which is impossible or extremely time-consuming to replicate based on cursory description of the methodology. Is this methodology any better than Mannian principal components? I don’t know. Josh RFTR Halpern/Eli Rabett is busy asserting that we should we trust these adjustments. But without any attribution to known statistical literature, it’s hard for me to simply take these adjustments on trust.

    I’ve got most of the papers – which ones are you looking for?

  31. BarryW
    Posted Aug 8, 2007 at 11:58 AM | Permalink

    Re 30

    Thanks for the offer. The three papers that are related to the undocumented changepoints are the ones I’m interested in. Especially the Lund and Wang since they deal with changepoints in trends. This is what they’re planning on using in V2 so I’d like to get a head start on understanding what we’re going to see.

    Alexandersson, H. and A. Moberg, 1997: Homogenization of Swedish temperature data. Part I: Homogeneity test for linear trends. Int. J. Climatol., 17, 25-34.

    Wang, X.L., 2003: Comments on “Detection of undocumented changepoints: A revision of the two-phase model”. J. Climate, 16, 3383-3385.

    Lund, R., and J. Reeves, 2002: Detection of undocumented changepoints: a revision of the two-phase regression model. J. Climate, 15, 2547-2554.

%d bloggers like this: