I wonder how this dedicated weather observer feels about having his readings adjusted by NCDC?

In my travels surveying weather stations around the United States, I met many dedicated observers like this one. It is sad indeed that their painstakingly recorded data by observers like this one gets adjusted by NCDC to give results that aren’t the same as what they observed. I have some comments, data, and photos about the station that follow, but let me say to Mr. Hendrickson first; thank you sincerely for your service and dedication.

Richard G. Hendrickson taking weather observations at his farm in Bridgehampton, New York. Photo: NOAA, 2008.NOAA honors New York farmer for 84 years of service as volunteer weather observer (press release)

When Richard G. Hendrickson (seen at right) logged his first weather observation for the U.S. Weather Bureau, the precursor to NOAA’s National Weather Service, Herbert Hoover occupied the White House. Since then the Bridgehampton, New York, farmer has filed twice daily reports, tallying more than 150,000 individual weather observations – playing a critical role in building our nation’s climate history.

As part of the National Weather Service Cooperative Observer Program, Hendrickson collects data from the weather observing station on his farm and calls in his observations – temperature, precipitation, wind and any other significant weather factors – to the weather service.

On July 27, Hendrickson, age 101, will receive an award for his long standing service – 84 years – to the nation. Since Hendrickson is first in the history of the program to serve for more than eight decades, the new 80-year service award will be named in his honor.

“Volunteer observers are the bedrock of weather data collection,” said I. Ross Dickman, meteorologist-in-charge of the New York weather forecast office. “Richard has contributed thousands of weather measurements to build the climate record for Long Island, and after 84 years, holds the title of the nation’s longest-serving volunteer weather observer. With this award, we honor Richard for his selfless dedication to his community and the country.”

Hendrickson started volunteering as a weather observer when he was 18 years old. His lifelong commitment stems from personal interest in weather and a sense of patriotism. “I enjoy observing the weather, it’s what I do for my country,” he said.

Hendrickson’s enthusiasm for weather extends beyond collecting data. In 1996 he authored, Winds of the Fish’s Tail, which highlights his years of observing the weather on Long Island’s east end. Hendrickson also writes a column on weather that is published in two eastern Long Island newspapers.

The award presentation will take place before an open house at the weather forecast office in Upton, New York. Throughout the day; residents are invited to tour the forecast operations floor, meet meteorologists and learn how forecasters track storms and issue warnings. The open house is an opportunity for the public to learn how to become weather-ready, become a storm spotter and see a weather balloon launch.

AWARD PRESENTATION:

Sunday, July 27, 9:45 a.m. to 10 a.m. EDT

New York Weather Forecast Office

175 Brookhaven Avenue, Upton, NY 11973

NOTE: Media must register with Tim Morrin to attend the ceremony, 631-924-0227

The National Weather Service’s Cooperative Observer Program has given scientists and researchers continuous observational data since the program’s inception more than a century ago. Today, over 8,700 volunteer observers participate in the nationwide program to provide daily reports on temperature, precipitation, and other weather factors such as snow depth, river levels and soil temperature. Long and continuous weather records provide an accurate picture of a locale’s normal weather and give climatologists a basis for predicting future trends. These data are invaluable for scientists studying floods, droughts, and heat and cold waves.

The first extensive network of cooperative stations was set up in the 1890s as a result of a Congressional Act that established the U.S. Weather Bureau. Many historic figures maintained weather records, including Benjamin Franklin, George Washington and Thomas Jefferson. Jefferson maintained an almost unbroken record of weather observations between 1776 and 1816, and Washington took weather observations just a few days before he died.

The National Weather Service New York forecast office located in Upton, New York, is the primary source of weather data, forecasts and warnings for about 18.6 million people in southeast New York, northeast New Jersey and southern Connecticut. Visit us at weather.gov/nyc and join us on Facebook and Twitter. For more on how to become weather-ready, visit Weather-Ready Nation.
NOAA’s mission is to understand and predict changes in the Earth’s environment, from the depths of the ocean to the surface of the sun, and to conserve and manage our coastal and marine resources.

==============================================================

First a look at the station itself from above. The coordinates are the ones given in NCDC’s HOMR metadata.

Bridgehampton_USHCN

The Stevenson Screen (white box between the row of trees and the house) is about 25 feet from the asphalt driveway, would would make it a Class 4 station, unacceptably sited:

Climate Reference Network Rating Guide – adopted from NCDC Climate Reference Network Handbook, 2002, specifications for siting (section 2.2.1) of NOAA’s new Climate Reference Network:

Class 1 (CRN1)- Flat and horizontal ground surrounded by a clear surface with a slope below 1/3 (<19deg). Grass/low vegetation ground cover <10 centimeters high. Sensors located at least 100 meters from artificial heating or reflecting surfaces, such as buildings, concrete surfaces, and parking lots. Far from large bodies of water, except if it is representative of the area, and then located at least 100 meters away. No shading when the sun elevation >3 degrees.

Class 2 (CRN2) – Same as Class 1 with the following differences. Surrounding Vegetation <25 centimeters. No artificial heating sources within 30m. No shading for a sun elevation >5deg.

Class 3 (CRN3) (error >=1C) – Same as Class 2, except no artificial heating sources within 10 meters.

Class 4 (CRN4) (error >= 2C) – Artificial heating sources <10 meters.

Class 5 (CRN5) (error >= 5C) – Temperature sensor located next to/above an artificial heating source, such a building, roof top, parking lot, or concrete surface.”

That’s not the fault of the observer, Mr. Hendrickson is working with what he has. NOAA/NWS actually installed and placed the station, and is responsible for its maintenance. The station is also boxed in by vegetation on three sides, along with the house for the fourth side, making it warmer than it should be due to wind inhibition.

What is even more interesting though is what happened to the data in 2012, according to this plot from NASA GISS of the station, there was quite a spike.

Bridgehampton_station_plot

Yet amazingly, even though Mr. Hendrickson has been dutifully reporting the daily data, and it is up to date, as seen in his May report below…

Bridgehampton_B91_May2014

…NASA GISS run by Gavin Schmidt, can’t seem to find the time to get their data set current for Bridgehampton, as seen here, only going to 2012. You’d think Gavin could tear himself away from Twitter long enough to at least get the data updated, especially since this man is so dedicated to the task.

More on all this in a later post.

UPDATE: 7/24/14 9AM I sent a Tweet yesterday to Gavin asking why Bridgehamptoon has not been updated at GISS since 2012, and as far as I know there has been no response.

Nick Stokes in comments thought that the lack of GISS updating was a GHCN problem, not a GISS problem.

I also asked the BEST team (who also use GHCN) and Zeke Hausfather responded almost immediately:

Looks up to date to me, as a file was just compiled this morning and is available up on the FTP site: ftp://ftp.ncdc.noaa.gov/pub/data/ghcn/v3/ . I believe they have a system that automatically compiles it daily.

Here is a chart of station observations in GHCN by month from today’s file. There are 2297 stations reporting so far for June 2014 (none for July, obviously, since its not over yet). If you check GHCN-daily instead of GHCN-monthly, you will find much more data from last month.

GHCN_station_reporting_count

My thanks to Zeke for that.

Also of interest are these notes in the status file for GHCN:

GHCNM, V3, status file (users can use this file to determine overall current status, including information related to previous changes and errata). ******************************************************************************** 07/14/2014 On or around 06/06/2014, there was an ingest problem with the “C” source data, (unpublished MCDW), and this caused a signficant reduction of data from that source. However, much of the data were still available through an alternate source (UK Met Office, “K” source flag). The ingest problem was resolved on 07/10/2014, and the expected frequency of “C” source data was restored. ******************************************************************************** 10/17/2013 Government operations have been restored, and regular monitoring of GHCN-Monthly will now resume. During the shutdown of government operations, some ingest of recent international data were not received. These data should be restored with the next processing cycle (e.g. 10/18/2013). ******************************************************************************** 10/01/2013 During the shutdown of government operations, GHCN-Monthly will continue to update automatically, but will not be monitored by the GHCN-Monthly team. We will also be unable to answer questions submitted to NCDC.GHCNM@noaa.gov until after government operations resume. ********************************************************************************

GHCN even continued to update during the “government shutdown” last year, and there is no note indicating late data for all of 2013.

So much for the Nick Stokes theory as to why GISS has not updated Bridgehampton. Now it’s back to Gavin and GISS.

I’m time limited for the next two days, so my promised update won’t happen until this weekend. Tony Heller has done some work in the meantime worth looking at here: http://stevengoddard.wordpress.com/2014/07/24/more-from-bridgehampton-ny/

The graph of adjustments show Bridgehampton’s data has been dramatically cooled in the past by as much as 1.5°F:

ScreenHunter_1270 Jul. 23 22.00

I have not double checked the graph above, but the spike at Bridgehampton in 2012 seems spurious, as I originally noted.

More on all this Saturday or Sunday when time permits.

About these ads

129 thoughts on “I wonder how this dedicated weather observer feels about having his readings adjusted by NCDC?

  1. Presumably the trees and the house have been there for a while, so even though the siting is not great, the results should be consistent.

  2. Hats off, a crisp salute, and a firm handshake to Mr Hendrickson!!! Having done what he did (at a Class 5 station no less) for just a short while, I can be appropriately awed by his dedication.

    Tenacity is a wondrous thing!!!

  3. More on all this in a later post….

    Please don’t tell me this man has been doing this fo 84 years….and they’ve been infilling his data

  4. ‘Hats Off!’ to Mr. Hendrickson, a man of dedication, integrity, and self respect.

    ‘Old School’, indeed!

  5. I never get tired of this quote:

    “If I have seen further it is by standing on the shoulders of giants.”

  6. Hats off to Mr. Richard G. Hendrickson.

    Tremendous amount of dedication exhibiting the American spirit.

  7. USHCN includes this station to the present day. May 2014 has an average temp of 14.41C in the USHCN “data”, which converts to about 57.9F. The average posted in your graphic….57.5F. Go figure…

  8. You can find month and annual averages here if it is accesible from non-government websites: http://xmacis.nrcc.cornell.edu/OKX/ and look for Bridgehampton. 2012 was very warm especially July, but 2013 turned out only 0.1 above the stations historical average. 2014 is not complete, but if the rest of the year is just average, the annual temperature average will end up well below the historical average.

    REPLY: That’s a link with ACCESS DENIED attached – Anthony

  9. The graph looks fairly flat, possibly even a slight downward trend, until about 1970 and then starts a steady upwards climb.

    It would be interesting to see what changes were made to the local environment since that time.

    When was the asphalt driveway put in?
    When were the screening trees planted?
    What are the adjustments over the reporting period?

    Hopefully Anthony’s “More on all this….” will explain.

  10. That is an inspirational record of service Mr. Hendrickson; and for all its shortcomings, our temperature records are phenomenal. They are easily obtained online so we can examine them, discuss them and even make withering criticism with no fear of reprisal. We live in a great country. We need to keep it great. Mr. Hendrickson certainly has done his part!

  11. Anthony,

    GISS supposedly uses the adjusted GHCN data. While the raw data file is fully up to date in GHCN, their adjusted file shows no readings past September 2012. They don’t even include a line in the file for the years 2013-2014. None of the raw data is flagged as erroneous, so I see no reason for leaving the adjusted data incomplete. They have done the same thing to many stations. One would think they’d leave some indication if they found defects while making the adjustments.

    TedL,

    The GISS data page isn’t available unless the graphic has been recently called up. It is only temporarily generated and available for a limited period of time.

  12. Data collection is an issue. Recent “adjustments” are made in the wrong direction. But what do I know?

  13. Double kudos. To farmer Hendrickson for his awesome persistance. To AW for ferreting it out.

    And double negatives to Gavin and gang for not even bothering to keep his records up to date.
    But then, Gavin, you probably thought your feeble efforts were good enough for government work.
    The OBummer part is, they were. Which is why in the real world you would have been fired long ago, and looking for minimum wage jobs requiring no skills. Since you evidently have none.
    Now I finally understand why OBummer wants to raise the minimum wage even if it increases unemployment. He wants to protect his borgs, like anyone associated with this embarrassment.

  14. As Bob Koss says above, this isn’t a GISS issue. They use only GHCN adjusted data. Here is the GHCN page for Bridgehampton. As you can see, they have unadjusted data up to date, but adjusted only to September 2012. Evidently something has been blocking the adjustment algorithm.

    REPLY:
    Actually if GISS were on the ball they’d be asking why the GHCN data isn’t updated yet, but…they don’t. At one time GISS imported USHCN and GHCN separately IIRC.

    In any event, just one more indication that the surface temperature record is a mess – Anthony

  15. Nick Stokes says: July 23, 2014 at 7:09 pm
    …. As you can see, they have unadjusted data up to date, but adjusted only to September 2012. Evidently something has been blocking the adjustment algorithm.

    Conscience?

    A well earned thank you and a hat tip to Mr Hendrickson.

  16. If Mr. Hendrickson took a reading twice a day, without missing a single one as he claims, then why are there so many 999.9 entries for the whole month?
    year 999.9’s
    1939 11
    1940 8
    1949 4
    1950 12
    1951 4
    1985 1
    1989 1
    2002 11
    2003 10
    2012 3
    The record stops in December 2012

  17. Maybe WUWT should start an annual award for dedication to Climate Science Integrity named in honor, “The Hendrickson Award”.

    What say yous?

  18. If only he had known that in 2014 there would be computer models that could generate all the data we need he could have saved himself a lot of bother…

  19. Eighty fours years? Why back in the 1990s it was only 70 something years. I see a trend. And kudos to farmer Hendrickson.

  20. Nick Stokes,

    your link, ftp://ftp.ncdc.noaa.gov/pub/data/ghcn/v3/products/stnplots/4/42500300889.gif ,
    shows that the pre-1980 recorded observations have been adjusted to colder temperatures, and the post-1980 recorded observations have been adjusted to warmer temperatures.

    These have been observed and recorded by the same observer for the last 80 years!

    What evidence about this observation site or what evidence about this observer requires these changes of the recorded data? Specifically.

  21. Nick Stokes,

    Absent any location or observer specific reasons for the GHCN adjustment of the recorded data from this observation site, the GHCN adjustments are just destruction of observation data.

  22. Destruction of recorded data, or ‘constructive alteration’ of recorded data; it is the same thing. Deception.

  23. jim says: July 23, 2014 at 9:08 pm
    “Absent any location or observer specific reasons for the GHCN adjustment of the recorded data from this observation site, the GHCN adjustments are just destruction of observation data.”

    Why not try to find out, then?

    The first thing you’ll find is that data is undestroyed. In fact, it is graphed in the page you refer to, which shows what is on the unadjusted file. And as the head post indicates, you can get the original docs on line.

    But in fact if you look at the adjustment history, there is just one sustained change in the early 1980’s. And sure enough, the metadata tells you there was a station move around that time.

  24. Dennis Kuzara,

    Those missing months of data are due to the adjustment algorithm GHCN employs. In the unadjusted(raw) file his record is perfect since 1931. No missing months.

  25. Alan says:
    July 23, 2014 at 4:16 pm

    “Presumably the trees and the house have been there for a while, so even though the siting is not great, the results should be consistent.”

    Trees and vegetation grow, air conditioners became more common, regional climate conditions could have changed…there is no such thing as stasis. It would be interesting to see pictures of the area from past years. Looking at the chart I would expect that a degree increase would reflect these changes since the 60s, and would have nothing to do with an increase due to AGW.

  26. “The Stevenson Screen (white box between the row of trees and the house) is about 25 feet from the asphalt driveway, would would make it a Class 4 station, unacceptably sited:”

    I remember visiting the campus of the University of Oklahoma one time way back around 1989 and quite distinctly remember seeing their weather observing station on the grounds. It was within 10 or 20 meters of a building which I believe might have housed the meteorology department there. This would make it a Class 4 station. It’s possible that the station might have been relocated since that time to a place that meets their guidelines a little closer.

  27. It bugs me that they seem to have no fear. That they can just keep doing what they are doing and there won’t be any repercussions. Some day this has to change.

    From the NCDC current monthly global surface temperature record here (not advertised but I thought some might want to know where the NCDC monthly anomalies are now saved – in the 1971-2000 base period versus the 1901-2000 or the 1981-2010 base period they commonly quote from).

    ftp://ftp.ncdc.noaa.gov/pub/data/mlost/operational/products

    To the archived data only going back to September 2011 (all other copies on the internet have been deleted).

    ftp://ftp.ncdc.noaa.gov/pub/data/mlost/archive/v3b/products/

    They have added about 0.08C to the trend. In just 2.5 years. And they started on the adjustment path about 1989, 25 years ago. 2.5 years = 0.08C, 25 years = ?.

    ——–

    The co-op volunteers and the regular observers have been betrayed.

  28. Nick Stokes,

    You should be proud of the work that you have done and all that you have written about climate, on the internet.

    So if some guy called Rick Rokes comes alone in the future, and says “The internet made a mistake, some letters were observed incorrectly, so when the letters “Nick Stokes” were recorded on the internet, those recorded letters were actually systemic erroneous observations of the letters “Rick Rokes”.

    So,in remedy of that systematic error, hence forth in all of the internet where the letters “Nick Stokes” are recorded, those letters will be changed to “Rick Rokes”. So the past will be correctly recorded for the future internet users. Past observations need to be corrected for comparison to the measure stick of the future present. GHCN style.

  29. Nick, please forget the cant that I posted last. Sorry. I went over the top, it’s not personal…

    Metadata. Station move. That lowers past temperature observations how or why? Site specifically.

  30. Nick,

    Site move can lower or raise corrections of the observational record. Why does the “station move” indicate that this observational record should be corrected to higher ‘observed’ temperatures? What specifically happened?

    Do you know (can you explain) anything more specific than some BEST style jackknife algorithm?

    I’ll take back everything that I wrote, if you can specifically show the quantified reason for the adjustment.

  31. Mr. Hendrickson is quoted saying, “We have polluted the stratosphere and because of that we have had warmer weather in the summer and milder weather in the winter and the potential of having heavy precipitation in the summer time increases– if not more rains, maybe they will be a little heavier than they have been in the past – you’ll notice your basement floods a little easier, your roof might leak a bit. We are in a period in the cycle of global warming. We have polluted our stratosphere with our big factories and it will happen.”
    NOAA says Mr. Hendrickson’s Bridgehampton weather station observation times weren’t recorded until the late 1940s. Then up until May 2008, the observations were taken at 8:00 pm. From then till now they were taken at 8:00 am. Mr. Hendrickson probably would not mind the data being adjusted to allow for the change in time of observation.

  32. Back in 1995 I was asked by specialists who had actual reading for large area of Sweden to forward those to scientists at Tema, Linköpings University. Those so called scientists told me they didn’t need correct readings – it was easier to interpolate and/or extrapolate in computers……

  33. Nick,

    It is not obvious from your link that there was an actual change of the station location. The measurement of the station’s location changed. Can we ask the hundred year old gentleman to find the facts of any station relocation? (Before you put me down, did you know the actual locations of the station? Do you know it/them now? I don’t. So you’re right, if it did move, and if the move changed the observational temperatures that were measured.)

  34. jim says: July 23, 2014 at 10:42 pm
    “Site move can lower or raise corrections of the observational record. Why does the “station move” indicate that this observational record should be corrected to higher ‘observed’ temperatures? What specifically happened?”

    According to the location history, on 17 July 1985, the station moved 0.1 miles N. According to the equipment history, on the same day an MMTS system was installed. According to another note I saw, the CRS system was kept as a backup.

    The change to MMTS itself requires an adjustment, which usually has warming effect. As to the relocation, it seems that the MMTS is further from the asphalt mentioned in the head post as a warming factor. If that is important, it means that the older readings were warmer than they should have been, relative to later.

    But as I said, you could try to find some of this out yourself.

  35. Avery Harden says:
    July 23, 2014 at 10:53 pm

    And yet according to the story he filed twice daily reports.

  36. Nick,

    I retract all that I wrote above. You are correct.

    But RE the station location history, the measurement of the location moved. Do you know if the actual location moved 160 m? That is not stated. I hope you don’t think that the MMTS is the yellow spot in the areal photo at the top of the post. In the tilled field.

  37. Nick,

    From the areal photo, its hard to see how a move of 160 m could stay on his property, with out regard to asphalt. So where is the MMTS located now, relative to the old Stephenson Screen? Is it to be seen on google earth. The more I look, the less that I believe you.

    Can we ask the gentleman? There is a difference between field work and desk work.

  38. Nick,

    You said “…on 17 July 1985, the station moved 0.1 miles N. ”

    Do you have any factual evidence of your statement?

    Looking at the photo at the top of the post, the indicated station relocation is impossible. The new station location is two neighbors parcels to the north.

    Do you know anything, or are you guessing from a desk?

  39. jim says: July 24, 2014 at 12:00 am
    “You said “…on 17 July 1985, the station moved 0.1 miles N. ”
    Do you have any factual evidence of your statement?”

    I said “According to the location history, on 17 July 1985, the station moved 0.1 miles N.”. There is a link there. That is what it says.

    I didn’t personally implement this move. I’m simply pointing to what is recorded. I personally think the 0.1 miles might be an approximation.

  40. Nick,

    “But as I said, you could try to find some of this out yourself.”

    Ok.

    “‘According to another note I saw, the CRS system was kept as a backup.”

    Where is that note?

  41. @ NIck Stokes

    I am not being snarky-
    ‘The first GPS satellite was launched in 1978 and the first products for civilian consumers appeared in the mid 1980’s. It was in 1984 that President Reagan announced that a portion of the capabilities of GPS would be made availabe to the civil community.’

    http://infohost.nmt.edu/~mreece/gps/history.html

    Changing times.

  42. https://www.ncdc.noaa.gov/homr/#ncdcstnid=10500064&tab=MSHR

    I see two station ‘moves’ that appear to be nothing more than a change in the precision of the way the lon / lat data is logged.

    Was the 59ft to 60ft ( above sea level ) change a physical move or not? It seems to coincide with some administrative changes at the same time.

    It will be interesting to see why the “corrected” data ends up with significant breaks when the original data was perfectly continuous.

    I look forward to the future article we are promised on this one.

    http://www.ncdc.noaa.gov/cdo-web/datasets/GHCND/stations/GHCND:USC00300889/detail

    Links on that page don’t seem to work. Does it require Java or something ridiculous to deliver a text file? WTF?

  43. Nick,

    You are excusing or explaining the appropriateness of the changing of the recorded data.

    I’m asking why the quantitative changes are appropriate.

    You’re saying to me “Find out for your self.”

    You provide links.

    The links don’t support what you contend.

    I ask for specific quanta.

    You say “Find out for your self”

    Wash. Rinse. Repeat.

  44. Having worked at a major data analytics company, apart from Mr Hendrickson’s original data collection, the onward handling processes and QA look very amateurish at best, would never pass any sort of audit, and would not attract commercial customers – to unreliable, and I would echo the comment that Gavin Schmidt would never get a job in the real economy based on this (non) performance.

  45. By eye it looks like the 1979-1981 ‘correction’ was about 0.35 deg C. If it was an equipment change is should happen from one day to the next , not spread over 2 years.

    I thought MMTS correction was 0.1 not 0.35 ??

    There was also a similar ‘correction’ from 1944-1946, what was that about?

  46. You can see the raw and adjusted data at the GHCN ftp site (link in next comment).
    The raw data shows a warming of about 0.5C, and the adjusted data turn this into warming of about 1.6C.

    The adjusters at GHCN have also decided to throw away chunks of Mr Hendrickson’s data, around the years 1940, 1950 and 2000, leading to the gaps in the GISS graph shown in Anthony’s post.

  47. What do they do about screen maintenance?

    1. Over the years, the screen becomes weathered, and as it does so, it absorbs more heat. So how often are screens painted, and what adjustment is made over the period between painting to take out the gradually articifical warming trend caused by screen degradation?
    2. When repainted, does the new paint have the same absorption and reflective qualities as the old. Change in pigments could alter the heat patterns.
    3. How often are the screened cleaned? There will be gradual build up of dirt which also will impact upon the screen’s absorption and reflective qualities, very probably leading to a warming bias beween cleans.
    4. Rainfall will also impact upon the absorption and reflective characteristics of the screen since water has its own absorption/reflective characteristics. So how does this pan out when there are changes in monthly patterns of rainfall from year to year?
    5. Where I am, rainfall is often far from cleansing. I presently reside in a dry area, and rain often contains considerable quantities of brown/red sand. Normally, one has to clean the car and outside furniture after it rains (some months are worse than other, althoughh I think there has only been 2 or 3 days of rainfall this year so not a monthlu occurence). this impacts upon albedo.
    6. What adjustmenst are made when the screen is covered by snow?

    No doubt the surface station project goes into this in more detail, but the role the screen plays is oftne frequently over looked, and nearly all ‘issues’ lead to a slight warming bias such that some cooling ‘adjustment’ needs to be made to reflect issues with the screen

  48. Nick,

    “I didn’t personally implement this move. I’m simply pointing to what is recorded. I personally think the 0.1 miles might be an approximation.”

    So if the station location is a quanta with an unknown error range, maybe the corrections for the station temperature observations are quanta with unknown values. The unknown error of location is comparable to unknown error of temperature measurement, since the systematic error of temperature measurement is ’caused’ by the the systematic change of station location.

    So if the location change of the station is actually found to be small, the changes to the temperature data record should be also small too. No? Or the GHCN temperature data corrections might be an approximation of 0.1 miles?

  49. Nick,

    You are the one defending the alteration of the data, for the GHCN record. The burden of proof lies on the people who want to change the measured observations, for historical records.

  50. It seems to me that the relocation does not involve any change in the ‘correction’ data, so there may not be too much value in making a meal out of it.

    The main change in diff is centred on 1980 but affects one year either side.

    If that is supposed to be an equipment change there’s a problem. If there was not equipment change there’s a problem.

  51. Nick says: “According to another note I saw, the CRS system was kept as a backup.”

    That would be a valuable cross-check if it can be accessed.

    Since the MMTS change is documented as 1985, why the +0.35 ‘correction’ from 1979-1981 and not the expected MMTS adjustment in 1985 ??

  52. BTW, I should add thanks to Mr. Hendrickson for his outstanding dedication to the task. One of the things that is lacking in most available weather records when trying to use them for climate, is continuity. That makes his work especially valuable.

    Heartfelt thanks. May he have the good health and energy to maintain his records for many years to come ;)

  53. Sorry, I see Nick and Jim have already given the link to the GHCN ftp page for Bridgehampton.

    Another thing is that the adjusted data as presented now (July 2014) is significantly different from the adjusted data presented last month (June 2014) when I last downloaded the data. For example, in June, there was no gap in the data around 2002. But now, almost 2 years of data around 2002-3 have been deleted in the adjusted data and hence in the GISS graph.

    I will post some numbers later.
    This ‘feature’ is not unusual, it’s quite common in the adjustment process and has been noted at Paul Homewood’s blog.

  54. Here is the GHCN/NOAA/NCDC adjusted data for Bridgehampton in 1931, as downloaded a month ago, file ghcnm.tavg.v3.2.2.20140622.qca.dat

    425003008891931TAVG -115 U -114 U 241 U 717 U 1193 U 1767 U 2132 U 2121 U 1878 U 1307 U 819 U 303 U

    and here it is downloaded today, file ghcnm.tavg.v3.2.2.20140722.qca.dat

    425003008891931TAVG -81 U -80 U 276 U 751 U 1227 U 1801 U 2167 U 2155 U 1912 U 1341 U 854 U 337 U

    So since June, they have changed their minds about what the temperature was in 1931 by 0.34C.

    For comparison, here is the unadjusted data:
    425003008891931TAVG 1 U 18 U 351 U 839 U 1317 U 1872 U 2239 U 2234 U 1988 U 1428 U 946 U 412 U
    So in June 2014 they cooled 1931 by about 1.1C, and in July 2014 they cooled it by about 0.8C.

  55. Here’s the adjusted data for 2002 and 2003, as downloaded in June:

    425003008892002TAVG 212 U 208b U 463 U 995a U 1275b U 1854 U 2303 U 2330 U 1919b U 1207b U 682b U 99 U
    425003008892003TAVG -342 U -259 U 300a U 704 U 1218a U 1813 U 2227 U 2324 U 1870 U 1121b U 801 U 209 U

    and here it is downloaded today:

    425003008892002TAVG 213 U-9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999
    425003008892003TAVG-9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 -9999 802 U 210 U

    So last month the adjustment algorithm thought that the 2002 data was OK after adjustment, but today it thinks almost all the 2002-3 data should be discarded.

  56. “So in June 2014 they cooled 1931 by about 1.1C, and in July 2014 they cooled it by about 0.8C.”

    Must be the first time they’ve warmed the past instead of cooling it !

    Perhaps, despite saying that everything was “working as designed” they have been reviewing the adjustment algo and making some changes. If they haven’t and the algo can produce results that bounce around like that “it’s worse than we thought”.

    Could you check what I assume were supposed to be URLs for the two files. “here is…” gives nothing.

  57. ~Paul: Sorry, maybe they were not supposed to be URLs, could you drop those somewhere ( at least the June archived file )?

  58. i look forward to the next post . since tony heller began highlighting adjustment issues i had a feeling there was a schidt storm brewing,this would appear to confirm it. here we have an adjustment for a station move that did not happen, it is likely as a result of using gps . how many other stations have an adjustment for station moves for the same reason.

    there is a continuous record of data being recorded in the evening until fairly recently , with all changes documented , so there are no unknowns in terms of tobs or equipment changes .so now an absolute comparison can be made between a raw data set with this information exact and the adjusted “data”.
    as i said,i look forward to the next post.

  59. Avery Harden July 23, 2014 at 10:53 pm says:

    Mr. Hendrickson is quoted saying, “We have polluted the stratosphere and because of that we have had warmer weather in the summer and milder weather in the winter and the potential of having heavy precipitation in the summer time increases– if not more rains, maybe they will be a little heavier than they have been in the past – you’ll notice your basement floods a little easier, your roof might leak a bit. We are in a period in the cycle of global warming. We have polluted our stratosphere with our big factories and it will happen.”

    NOAA says Mr. Hendrickson’s Bridgehampton weather station observation times weren’t recorded until the late 1940s. Then up until May 2008, the observations were taken at 8:00 pm. From then till now they were taken at 8:00 am. Mr. Hendrickson probably would not mind the data being adjusted to allow for the change in time of observation.

    Not if he is scientific. There is no room in science for that cavalier attitude to tampering with records that you just exhibited ( and if you think so, just imagine what fun others might have tampering with records related to you ). Either way you have no business putting words in his mouth. That warmist propaganda attributed to him is probably cherry picked out of something larger anyway.

    So what to do? How about this … the simple, logical and formerly favored method of the concept of scientific controls.

    Recreate sites today using the original equipment on the same locations where possible, studiously using photographs as a guide, trees, asphalt, everything. Most importantly take the temps at the same time as was done back then. According to Leif something similar is done with sunspots using early telescopes.

    One thing is for sure IMHO, if you think the appropriate action to correcting incompatibilities due to 8:00am versus 8:00pm is to alter the historical dataset, then you are not a scientist by any stretch of the imagination. What that makes you is the exact opposite of a scientist. Hear that Mosher? It is more akin to Stalin’s cadre of artists airbrushing out inconvenient truths in photographs. Not science.

    When someone says we cannot compare today to the historical past ( I wonder if this is true in England’s CET ) then the the logical thing is to fix the site and make it compatible to the historical record … unadjusted. Or close the site.

    Perhaps as part of the Siting Project those involved could identifies which sites ( if any ) are comparable to the past, what is the longest comparable recording period of time, and which sites could be retrofitted or resurrected from the dead. Apples to Apples you know. Not Apples to Cherries, or Oranges or Tangerines or fake Apples.

  60. Temperature has to be adjusted because of the shadowing effect from air craft. Anthony you might remember before 911 there was someone ,I can’t remember who tried for years to get the American government to stop all air traffic over America for a day , so he could collect all the data he had organised from people like the above Richard Hendrickson and then came the fall of the twin towers and America suspended all air traffic for a couple of days. Didn’t he get the data he required to prove he’s shadowing effect theory. I thought the shadowing effect was responsible for a 2 deg drop in temps. When the hot exhaust from a jet flying at the top of the freezing troposphere is ejected ,what we see is cold air chasing hot or negative charge (heat) attracting positive charge (cold). One thing I see is climate science likes to use the theory that hot chases cold . I don’t agree with this theory because , the first electromagnetic field line is the tropopause (cold positive charge) which is attracted to the 6000 C (hot negatively charged )core and the hotter stratopause, that’s why it’s held in place and the tropopause works north south and not up and down. Temperature = electric potential at work or to put it another way it depends how fast the electron is moving around the nucleus that determines it’s temperature. Heat creates fission (seperation) and cold creates fussion ( attraction) One other thing to remember is the troposphere is a sea of negatively charged electrons.

  61. The dedication and stability of a Richard Hendrickson stands in stark contrast to the data collection that has prevailed in most of the world over most of the last 160 years.

    0.6° C. over the last 160-odd years? Given the state of the climate/weather data gathering system, the adjustments ( cough, cough ) made by GISS, the urban heat island effect, Chinese ( cough, cough ) weather stations/data, Russian ( cough, cough ) weather stations/data, Sub-Saharan African ( cough, cough ) weather stations/data— among a multitude of other problems, that’s a rounding error— at best.

    When one stops to consider the reliability of the historic temperature records, one is left to wonder if we are kidding ourselves about our ability to gauge the extent to which current temperatures are or are not higher or lower.

    Do you really believe that Russian temperature records from, say, 1917-1950 are reliable?

    Do you honestly believe that Chinese temperature records from, say, 1913-1980 are reliable?

    Do you seriously believe that Sub-Saharan African temperatures from, say, 1850-2012 are accurate?

    I don’t.

  62. Greg, I don’t think they have changed the algorithm recently, because this kind of thing happens all the time. The algorithm is just unstable somehow – as you say, it bounces around. This is discussed on Paul Homewood’s blog somewhere.

    You can find the latest file, ghcnm.tavg.latest.qca.tar.gz by going to Nick/Jim’s ftp link and backing up to /ghcn/v3/
    The June file isn’t available on the web as the files are overwritten. Try this link, assuming you are happy with .tar.gz files

    https://docs.google.com/file/d/0B1fcBLZDFH7RRFVidmk3QzJXcDQ/edit?pli=1

  63. ilma630 says:
    July 24, 2014 at 12:49 am
    Having worked at a major data analytics company, apart from Mr Hendrickson’s original data collection, the onward handling processes and QA look very amateurish at best, would never pass any sort of audit, …
    —————————

    That is a very good point.

    Where is the audit trail?

    If anything, NCDC goes out of is way to erase any type of past tracking.

  64. “The June file isn’t available on the web as the files are overwritten. Try this link, assuming you are happy with .tar.gz files ”

    Thanks Paul, that’s fine.
    I now archive everything I get because data files seem to have a short half life these days. I never overwrite with an update, the first thing I do is a diff to see what has changed apart from appending new data.

  65. Nick Stokes says:

    July 23, 2014 at 11:15 pm

    The change to MMTS itself requires an adjustment, which usually has warming effect. As to the relocation, it seems that the MMTS is further from the asphalt mentioned in the head post as a warming factor. If that is important, it means that the older readings were warmer than they should have been, relative to later.

    Yes, and by that logic, all of the earlier readings should be adjusted cooler.

    I believe it has been noted that any incorrectly site – Not Class 1 (CRN1) – may give an artificially higher than correct reading.

    The complexities involved in attempting to properly correct each properly operating non-Class1 site with its unique, ever changing environment, is virtually impossible using current information gathering techniques.

  66. re: lee July 24, 2014 at 12:40 am

    No one here has probably used the old-timey paper USGS 7 1/2 minute topographical maps to ‘take’ long and lat measurements; I have, and it is NOT surprising that with the advent of GPS a more accurate position was recorded.

    Too many ppl today with only a lot of ‘desk’ experience, as poster ‘jim’ above alludes to.

    .

  67. re: lee on July 24, 2014 at 12:40 am
    ‘The first GPS satellite was launched in 1978 …

    A brief side note: We had at the time 3 (three) sats in orbit in 1978 as the fledgling GPS NAVSTAR project was still in its infancy, the hardware of the day still being developed at several companies in the US.

    Yuma Proving Grounds had a ‘ground range’ where aircraft could fly patterns while taking data on new hardware before the full constellation of sats was in place. I was with TI at the time working on their HDUE (High Dynamic User Equipment), performing ‘factory test’ of product and providing technical support to personnel performing environmental testing (shake and bake) on the HDUE ‘boxes’.

    Later, in the 90’s while at MetroCel Cellular (to be bought by ATT WS) I made use of GPS and USGS topo maps to do cellular system planning and site selection; with the GPS sats having SA (Selective Availability) turned on one’s position on earth could be seen to ‘drift’ +-300 feet or more as SA was meant to ‘dither’ the computed position about a centroid point. Not until sometime in Pres. Clinton’s term (after May 1, 2000) SA was turned off, resulting in a more accurate, instantaneous GPS position report.

    SA – http://www.gps.gov/systems/gps/modernization/sa/faq/

    GPS – http://en.wikipedia.org/wiki/Global_Positioning_System

    .

  68. Maybe I missed something, someone living near this gentleman should walk up and ask him in the decades you have been making readings, how far has that station moved? When was the asphalt driveway built? Again, if I missed this, my apologies. If it has not moved, screw you nick.

  69. re: jmorpuss July 24, 2014 at 3:29 am
    Temperature has to be adjusted because of the shadowing effect from air craft. Anthony you might remember before 911 there was someone ,I can’t remember who …
    One thing I see is climate science likes to use the theory that hot chases cold . I don’t agree with this theory because , the first electromagnetic field line is the tropopause (cold positive charge) which is attracted to the 6000 C (hot negatively charged )core and the hotter

    .

    You’re the Electric Universe guy right? I remember you (others should too) … off the wall ‘theories’ with no basis in known physics and all that …

    Trying to sneak in here under the RADAR?

    .

  70. @Dennis Kuzara et al. The GISS 2012 spike is due to missing data for Sep-Oct-Nov. The 12.0C average for 2012 is based on only the previous three quarters (3.1C, 11.0C, 22.0C), so naturally there’s a spike.
    I apologize if someone else pointed this out earlier.
    Regardless, Mr. Hendrickson is my hero.

  71. The question I would like answered was whether the gentleman himself compensated for TOB’s? or whether he just blindly logged the data from the min max bars at exactly the same time each day?

    The recorders that I talked to all compensated for exceptionally hot or cold periods. They were extremely careful to get the accurate temperatures for the day and understood the limitations of the device.

    Some of them did their own infilling though, taking temperatures from the Daily paper to fill holes in their record.

  72. He observed the temperature twice a day. That should make any TOPS adjustment unnecessary?

  73. The universal curse of Civil Service – dedication and professionalism will get your work noticed – eventually – but when there are political agendas afoot, your work will get no respect.

  74. Nick, you said

    “According to the location history, on 17 July 1985, the station moved 0.1 miles N. According to the equipment history, on the same day an MMTS system was installed. According to another note I saw, the CRS system was kept as a backup.

    The change to MMTS itself requires an adjustment, which usually has warming effect. As to the relocation, it seems that the MMTS is further from the asphalt mentioned in the head post as a warming factor. If that is important, it means that the older readings were warmer than they should have been, relative to later.”

    You need to go back and look at the map again. If the station was moved 0.1 mile N then the original location was 0.1 mile (~500 ft) S of the present location. According Google maps, that area is an open field, just north of a barn. So in effect, the station was moved from about a class 2 location to its present class 4 location. Cooler to warmer. Seems a cooling adjustment pre-move doesn’t track here.

    Avery Hardin,
    As Nick points out since 1985, the system has been an MMTS: Min – Max system. Min and Max temps do not depend on TOBS so a TOBS adjustment would not be appropriate either.

    I suspect that adjustments are bulk system adjustment made based on an assumption of application rather than the individual appropriateness as dictated by each site.

  75. I’ll have an update to this post this weekend (as a new post), right now I’m time limited by other duties.
    In the meantime I’ve made a short update to this one.

    Nick Stokes theory about lack of update at GISS being a GHCN problem doesn’t hold water.

  76. HELP

    I am not sure that I fully understand the TOB issue, and perhaps someone would kindly explain it to me, because my initial (uninformed) view is that it is a date of record, not a time of observation problem. Eg,

    IF on 6th June, at say 10 am, I observe and record the max/min recordings, what I am seeing is the max and minimum temperature for 5th June. Whereas,

    If on 6th June at say 6pm, I observe and record the max/min, what I am seeing is the max for 6th June, and the minimum for 5th June.

    Now if I observe temperaturs over a period of say 40 years, but at some stage say after 12 years, I change my habit from making my observations at 10 am but instead I make my observations at 6pm, I will have correctly recorded every maximum and every minimum temperature that that station has observed, it is just that i will have got one day out of sync.

    However, we are not concerned with making a comparison from one specific day on year X to the same specific day the following year. All we are after is the average max/min as observed by that station. The fact that one day’s data has got out of sync, makes no real difference to the average, and the resulting trend, and I would suspect that the ‘error’ therby created is rather less than trying to make some theoretical correction to what that day’s temperature would have been had it been taken at a different time, eg., when I make my observation at 10 am on 6th June, I know what the maximum temperature was the day before (because that is what the max is showing me), and I know the current 10 am temperature, but it is a complete guess to work out what the temperature will be at say 1pm or 2pm or 3pm later that day on 6th June. And when I come to check the temperature at 10 am on 7th june, I will know what the high for 6th June was.

    Now I can see that there may be problems if you keep on checking temperatures at random times, without routine, but when you take temperatures twice a day (if the times are reasonably straddled early morning and early evenning) you will always be getting a correct max/min and at worst you will merely be off-syncing an observation. I accept that very occassionally, a day when there are temperature inversions may lead to an error, but this will be rare, and again the ‘error’ is likely to be less than trying to make some guess as to what the high would be if observation was made at different times.

    To me TOB becomes an issue only when you get sparse collection of data, say on a few days one week, one day the next week, 5 days the following week, 1 day the week after etc etc, and/or when each and every day’s observation is made at completely random times,. But if that is the type of data coming from a station it should be thrown out, not homogenised.

    If you are getting data collected in the dilligent manner as shown by Richard G. Hendrickson, twice a day, I can’t see the need for a TOB adjustment. It seems to me that it is more likely to create an error, than to correct one.

    As i say, I am probably missing something important, and I would appreciate a simple explanation. THANKS.

  77. Further to my post above, Correction:

    IF on 6th June, at say 10 am, I observe and record the max/min recordings, what I am seeing is the max for 5th June, and the minimum temperature for 6th June. Whereas,

    If on 6th June at say 6pm, I observe and record the max/min, what I am seeing is the max for 6th June, and the minimum for 6th June.

  78. re: bwanajohn July 24, 2014 at 8:28 am

    You need to go back and look at the map again. If the station was moved 0.1 mile N then the original location was 0.1 mile (~500 ft) S of the present location. According Google maps, that area is an open field, just north of a barn. So in effect, the station was moved from about a class 2 location to its present class 4 location. Cooler to warmer. Seems a cooling adjustment pre-move doesn’t track here.

    Presumes original long/lat determination was accurate. I assert it was not. Later the position was (probably) corrected with a GPS-based lat long coordinate pair. The first-ever lat/long pair was probably determined by the county engineer (responsible for bridges and roadways) who had access to topo maps of the day via a description of the shelter position over the telephone.

    You’re at the limits of accuracy in determining the long/lat on a map designed for that purpose or if doing a ‘star site’/sextant-based reading even.

    .

  79. I’m trying to understand the logic involved in changing the historical data. The station moved, so they changed the data? Unacceptable. The data should have been left alone and the station marked as new and the old location marked as closed. I know that probably doesn’t fit into the agenda, but that’s the way it should be done.

  80. Will someone near by please go an interview this extraordinary gentleman?

    Ask him if the station has ever been moved. Ask when the driveway was installed. Ask if the home is original. Ask if the roofing on the house has changed over the years. Ask if he has continued to use white wash and when/if he switched to latex paint. Ask when the trees were planted and what was there before. Ask if air conditioning has been installed in the home. Ask if the color of the house has changed. Ask if there have been equipment changes. Ask how the yard is watered and how has that changed over time.

    This is a great opportunity to document how changes in local environment change readings. Bring out the tip jar, I’ll contribute to a travel fund. It will take a few days of visiting, listening to stories, looking at pictures….

  81. Nick,

    Looking at the information on the NCDC link and comparing the imputed moves with the stated fixes using Google Earth, the 0.1 mile move is highly unlikely to be true. There’s a large McMansion at that location and no indication of a suitable site for an MMTS station “north” of the location in the photo. Also the corrections to the lat-long data simply move the location from the house yard west into the ploughed field. There DOES appear to be an installation, possibly an MMTS (or a bird feeder – they would both cast similar shadows), about 12 meters southeast of the Stevenson Screen.

    It may be that regardless of what the NCDC record says, the NCDC information is not trustworthy.

  82. Forget when the tree’s were planted, a better question would be when was the field converted from forest? and does the wind blow from the SW most of the time ;0)

  83. richard verney says: July 24, 2014 at 9:35 am
    “As i say, I am probably missing something important, and I would appreciate a simple explanation. THANKS.”

    I’ve given an illustration here. I took hourly data over three years from Boulder Colorado, and shown the average that you would calculate if you were reading an accurate min/max thermometer on a prescribed daily schedule. TOBS makes a big difference. The reason is that reading in the afternoon means warm afternoons tend to be double counted. On a warm Monday, if you read at 5pm, it may still be warm enough to provide the Tuesday max as well. But you’ll only count each minimum once. Conversely, if you read in the morning, minima will be double counted.

  84. Duster says: July 24, 2014 at 11:11 am
    “It may be that regardless of what the NCDC record says, the NCDC information is not trustworthy.”

    It may indeed. It isn’t actually a NCDC record; they just have to work with what was recorded. FWIW, I agree with your 12m interpretation.

    It is partly because of the unreliability of metadata that NCDC tends to rely on homogenisation algorithms.

  85. Nick Stokes says:
    July 24, 2014 at 2:35 pm

    During the 19th century in California the Department of the Interior hired surveyors to survey what is known as the Public Land Survey. The maps are by and large pretty good and with access to the surveyor’s notes you can normally follow the original survey route, identifying trees rocks, creeks etc. as they occur in reality. But sometimes, it seems clear that the old surveyor was either lost from the get go, or sat in a bar and made every thing up. Fictitious streams and mountains cause me to lean toward the bar explanation. Looking at location information like that in the NCDC link has me wondering about whoever installed the MMTS.

    In looking at the location information it seems that 52.8 feet is 0.01 mile – about 16 meters, so maybe 0.1 miles is a transcription error. There would then be no significant “site move”. But “north”? Where was his compass? Another location problem is the elevation. The one-foot move is almost certainly spurious. There is not that much slope in the yard. The difference is in all probability due to the shift from the NAVD 1929 vertical datum to NAVD88. Since they use different geoids there will be a minor difference in elevation. Is homogenization necessary because the folks recording the metadata didn’t know what they were doing?

  86. I’m suspecting that there are a lot of questionable station ‘moves’, based on the gyrations my local weather station/river gauge has supposedly done since 1948…the problem is, the station/gauge was housed in a purpose built, stone structure anchored to the bedrock of the river bank…the current coordinates for my local station place it in a barn, on the top of a hill, about 1/4 mile from the river!

  87. Duster says: July 24, 2014 at 3:15 pm
    “Is homogenization necessary because the folks recording the metadata didn’t know what they were doing?”

    Imperfection of the record is part of it. But in fact, whether it was .1 or .01 mi, in which direction, is neither here nor there. None of that helps quantify the effect of the move. For that, the only usable information is in the record itself, and that is what the algorithm estimates. It also estimates the time of the move, which here does not exactly correspond to the metadata.

    The thing is, there was a move, and a change to MMTS, and it had a discernible effect on the record. It can’t be estimated perfectly, but it’s better to estimate than pretend it didn’t happen.

  88. “Nick Stokes says:
    July 24, 2014 at 3:33 pm

    The thing is, there was a move, and a change to MMTS, and it had a discernible effect on the record. It can’t be estimated perfectly, but it’s better to estimate than pretend it didn’t happen.”

    What you seem to be saying is that two adjustments are being applied to the data, one for a move and one for an equipment change.

    The question is, was the move ‘real’ or just a change in the location data. And without knowing if it was an actual physical move, then an adjustment for it may be unwarranted…but seems to be being applied anyway.

    Changing equipment is probably a legitimate reason for adjustment. Also, this case it seems the equipment change was not to replace broken equipment but to upgrade it, with the old equipment remaining on site. So wouldn’t it be better to get a side-by-side comparison as opposed to an estimate from an off-site source?

  89. Zeke is probably right from the BEST point of view, assuming that it is the unadjusted GHCNM data which they use, and which he has looked at today. But Gistemp uses the adjusted GHCNM data, and that has stopped at 2012 for some time, so Gavin and GISS are not to blame here.

    I’ve a recent post on GHCN issues relating to Irish stations, including this current data disappearance, at GHCN data collection issues (from an Irish perspective).
    The January 2014 data for Bridgehampton was present in the early to mid adjusted (qca) and unadjusted (qcu) GHCNM data, but went missing later in February in the adjusted data, showing data only to 2012. The following image shows the last year present in each file for various dates this year (I ran it this morning our time before today’s file became available, but I’ve checked that 2012 is the last year there too).

    There look like some other issues with Bridgehampton data as well, although not the same ones as I described in the post referred to above. I’ll add another comment here when I’ve reviewed them, with possibly a fuller post of my own if they need that rather than just a comment.

  90. mjc says: July 24, 2014 at 4:06 pm
    “What you seem to be saying is that two adjustments are being applied to the data, one for a move and one for an equipment change.”

    No, the change was applied based on the observed change in the record. They don’t try to allocate cause. GHCN is global, and while metadata in the US may be unreliable, elsewhere it is mostly inaccessible.

    However, here we do know that there was a change to MMTS and a physical move, at about the time that the algorithm found a discontinuity in the temperature. We know there was a physical move, because the old CRS is still there, and doesn’t seem to have a MMTS adjacent. It fact, as Duster says, it looks a lot like it is a few metres further along the building. But as I said above, if the asphalt really was a factor, then it has receded.

  91. ” Nick Stokes says:
    July 24, 2014 at 4:21 pm

    However, here we do know that there was a change to MMTS and a physical move, at about the time that the algorithm found a discontinuity in the temperature. We know there was a physical move, because the old CRS is still there, and doesn’t seem to have a MMTS adjacent. It fact, as Duster says, it looks a lot like it is a few metres further along the building. But as I said above, if the asphalt really was a factor, then it has receded.”

    So in other words…

    The change is NOT a correction factor, because it is not correcting for a specific problem, but rather, it is being made because there was an unexpected discontinuity in the data. Basically, this means that the amount being applied is really just an arbitrary number?

  92. Bridgehamptom NY…..Should not Mr. Hendrikson be a climate refugee by now?
    How can his temp apparatus still be standing after that once-in-a-lifetime super-duper-storm known as Sandy?
    Hasn’t LI been wiped off the map by tornadoes and drought already?

  93. On top of all that – -bring the First Lady in to tell this guy what he ought to be eating. I have a feeling he is a bacon-and-eggs type of guy, and he needs some soy and gluten-free instead.

  94. Nick Stokes says:
    July 24, 2014 at 2:29 pm

    richard verney says: July 24, 2014 at 9:35 am
    “As i say, I am probably missing something important, and I would appreciate a simple explanation. THANKS.”

    “I’ve given an illustration here. I took hourly data over three years from Boulder Colorado, and shown the average that you would calculate if you were reading an accurate min/max thermometer on a prescribed daily schedule. TOBS makes a big difference. The reason is that reading in the afternoon means warm afternoons tend to be double counted. On a warm Monday, if you read at 5pm, it may still be warm enough to provide the Tuesday max as well. But you’ll only count each minimum once. Conversely, if you read in the morning, minima will be double counted.”

    You mean Hendrickson scratched out his log entry and re-entered the minima in the morning?
    Anthony provided a shot of his log above. Do you think he would have been unaware of this?

  95. Great story about the man’s achievement.
    A very sad story about Man’s achievement.
    Never have I seen such blatant disregard for accepting responsibility as those that worship CAGW.
    The would rather change the data that doesn’t promote a cooling, hide the data that shows the cooling, or simply extend their predictions until no one remembers what they did.

    Brilliant !

  96. I’ve added two more issue images to that file, and changed the title accordingly to Images for WUWT comment

    2012 adjusted (qca) and unadjusted (qcu) data going missing, reverting back to 2008, with adjusted data first passing through 2011, reappearing sometime in January 2013.

    January 2012 value first recorded as 1.95°C, with code “a” indicating one day missing in calculation of monthly mean, then as 2.28°C with no days missing. The missing day would need a temperature of at least 12.03°C to effect this change, but TMAX for January 2012 is only 7.09°C, the maximum TMAX for January is 8.28°C in 1932 (February also changes at the same time with no indication of missing days, March loses a day, April has a “suspicious looking” 9.99°C with one day missing, which later changes to 10.13°C, still with one day missing).

  97. This wonderful man doing this observing for all this time would no doubt keep very accurate log books. His data is not a state secret, asking the man directly and getting access to his records should not be hard, he would be proud of what he has done. The entire 84 years of records would be a very interesting study, and point a finger directly at the data manipulators.

  98. On Sunday, Chetta and I are taking the 7:30AM Port Jeff ferry over to go to the ceremony. If you could ask Richard G. Hendrickson one question what would it be? If you could ask NCDC one question what would that be?

  99. My question would be: Did they ever calibrate the thermometers, especially after some changes?

  100. From http://www.noaanews.noaa.gov/stories2014/20140721_hendrickson.html:

    Since [1930], [Richard G. Hendrickson] has filed twice daily reports.

    In trying to verify when the reports were filed, I looked at the original forms B-91 available at http://www.ncdc.noaa.gov/IPS/coop/coop.html. From the forms B-91, I could not verify this statement. However, what I did find is that Mr. Hendrickson is identified as the observer only beginning in January of 1944, which would mean that he has only been an observer for 70 years according to the forms. Before that, Howard P. Corwith is identified as the observer from Sep. 1940 until Dec. 1943. Ernest S. Clowes is identified as the observer from the beginning of the record in Aug. 1930 until Aug. 1940.

    The observation times were, as was common then, in the afternoon, but they vary from as early as 4:15 PM to 8:00 PM. There did not seem to be a consistent observation time until Oct. 1956, when the observation time is noted as 8:00 PM until Mar. 2012, when it is noted as 7:00 AM. An MMTS was installed in Jul. 1985 with the Stevenson Screen and the liquid-in-glass thermometer remaining as a backup. There is a note in Jan. 1997 that the “OBSVN” was changed, but not from what to what. In Apr. 2008, there is a note that the “morning observation time” was changed to 8:00 AM “using tel[ephone]“, but not from what time. The note says that the observer would call in his info and that the office “puts data in wxcoder 3.” There is no other evidence of two observations a day, as the observation frequency is consistently reported as “DAILY.” Metadata was obtained at http://www.ncdc.noaa.gov/homr/, but the changes in the identified observer are not noted in the metadata. The only station relocation reported was at the time the MMTS was first installed in 1985: “0.1 mi N.” There does not appear to be a record of twice daily observations other than the cryptic note in 2008. I don’t know how a reasonable adjustment for the TOBS error could be made on this record. Following is comma delimited info that I put together in chronological order.

    DATE,OBS, TOBS (Form B-91),NOTE
    Aug-1930,Ernest S. Clowes,06:00 PM
    Dec-1936,Ernest S. Clowes,4:45 to 4:15 PM
    Jan-1939,Ernest S. Clowes,05:00 PM
    Jan-1939,Ernest S. Clowes,05:00 PM
    Jan-1940,Ernest S. Clowes,05:00 PM
    Jun-1940,Ernest S. Clowes,08:00 PM
    Aug-1940,Ernest S. Clowes,7:00 to 8:00 PM
    Sep-1940,Howard P. Corwith,05:00 PM
    Oct-1940,Howard P. Corwith,05:00 PM
    Feb-1941,H.P. Corwith,04:30 PM,Change of observer
    Mar-1941,H.P. Corwith,05:00 PM
    Jun-1941,H.P. Corwith,05:00 PM
    Dec-1941,H.P. Corwith,05:00 PM
    Jul-1942,H.P. Corwith,05:00 PM
    Jul-1943,H.P. Corwith,05:00 PM
    Dec-1943,H.P. Corwith,05:00 PM
    Jan-1944,Richard G. Hendrickson,05:00 PM,Change of observer
    Aug-1944,Richard G. Hendrickson,05:00 PM
    Jan-1945,Richard G. Hendrickson,05:00 PM
    Oct-1956,Richard G. Hendrickson,08:00 PM
    Jan-1964,Richard G. Hendrickson,08:00 PM
    Jul-1973,Richard G. Hendrickson,08:00 PM
    Nov-1983,Richard G. Hendrickson,08:00 PM
    17-Jul-1985,,,”Relocation .1 mi N MMTS INSTALLED, CRS/MXMN REMAIN AS BACK-UP”
    24-Jan-1997,,,”UPDATE EQUIPMENT, ADD MMTS, UPDATE GPS LAT/LON, REMOVE ROSA, CHANGE OBSVN AND EQUIP SECTIONS”
    Sep-1997,Richard G. Hendrickson,08:00 PM
    Oct-2005,Richard G. Hendrickson,08:00 PM
    Jun-2006,Richard G. Hendrickson,08:00 PM
    14-Apr-2008,,,AT THE EFFECTIVE DATE CHANGED MORNING REPORTING TIME TO 0800 USING TEL. OBSERVER CALLS DATA INTO REP OFFICE. REP OFFICE PUTS DATA INTO WXCODER 3.
    Dec-2008,Richard G. Hendrickson,08:00 PM
    Sep-2009,Richard G. Hendrickson,08:00 PM
    Dec-2011,Richard G. Hendrickson,08:00 PM
    Jan-2012,Richard G. Hendrickson,08:00 PM
    Feb-2012,Richard G. Hendrickson,08:00 PM,Last paper Form B-91
    Mar-2012,Richard G. Hendrickson,07:00 AM,First Digital Weathercoder 3 Form B-91

  101. The change of observer from Mr. Clowes to Mr. Corwith as incorrectly shown in the data as being in Feb. 1941. It was actually in Sep. 1940.

  102. Presumably the trees and the house have been there for a while, so even though the siting is not great, the results should be consistent.

    That is not what we find. We find that bad siting — consistent and unchanged over time — greatly increases the trend (sic). Warming or cooling trend is exaggerated, but since we have seen more warming than cooling, the microsite bias is on the side of warming.

Comments are closed.