Correcting and Calculating the Size of Adjustments in the USHCN
By Anthony Watts and Zeke Hausfather
A recent WUWT post included a figure which showed the difference between raw and fully adjusted data in the United States Historical Climatology Network (USHCN). The figure, used in that WUWT post was from from Steven Goddard’s website, and in addition to the delta from adjustments over the last century, included a large spike of over 1 degree F for the first three months of 2014. That spike struck some as unrealistic, but knowing that a lot of adjustment goes into producing the final temperature record, some weren’t surprised at all. This essay is about finding the true reason behind that spike.
One commenter on that WUWT thread, Chip Knappenberger, said he didn’t see anything amiss when plotting the same data in other ways, and wondered in an email to Anthony Watts if the spike was real or not.
Anthony replied to Knappenberger via email that he thought it was related to late data reporting, and later repeated the same comment in an email to Zeke Hausfather, while simultaneously posting it to Nick Stokes blog, who had also been looking into the spike.
This spike at the end may be related to the “late data” problem we see with GHCN/GISS and NCDC’s “state of the climate” reports. They publish the numbers ahead of dataset completeness, and they have warmer values, because I’m betting a lot of the rural stations come in later, by mail, rather than the weathercoder touch tone entries. Lot of older observers in USHCN, and I’ve met dozens. They don’t like the weathercoder touch-tone entry because they say it is easy to make mistakes.
And, having tried it myself a couple of times, and being a young agile whippersnapper, I screw it up too.
The USHCN data seems to show completed data where there is no corresponding raw monthly station data (since it isn’t in yet) which may be generated by infilling/processing….resulting in that spike. Or it could be a bug in Goddard’s coding of some sorts. I just don’t see it since I have the code. I’ve given it to Zeke to see what he makes of it.
Yes the USHCN 1 and USHCN 2.5 have different processes, resulting in different offsets. The one thing common to all of it though is that it cools the past, and many people don’t see that as a justifiable or even an honest adjustment.
It may shrink as monthly values come in.
Watts had asked Goddard for his code to reproduce that plot, and he kindly provided it. It consists of a C++ program to ingest the USHCN raw and finalized data and average it to create annual values, plus an Excel spreadsheet to compare the two resultant data sets. Upon first inspection, Watts couldn’t see anything obviously wrong with it, nor could Knappenberger. Watts also shared the code with Hausfather.
After Watts sent the email to him regarding the late reporting issue, Hausfather investigated that idea, and ran some different tests and created plots which demonstrate how the spike was created due to that late reporting problem. Stokes came to the same conclusion after Watts’ comment on his blog.
Hausfather, in the email exchange with Watts on the reporting issue wrote:
Goddard appears just to average all the stations readings for each year in each dataset, which will cause issues since you aren’t converting things into anomalies or doing any sort of gridding/spatial weighting. I suspect the remaining difference between his results and those of Nick/myself are due to that. Not using anomalies would also explain the spike, as some stations not reporting could significantly skew absolute temps because of baseline differences due to elevation, etc.”
From that discussion came the idea to do this joint essay.
To figure out the best way to estimate the effect of adjustments, we look at four difference methods:
1. The All Absolute Approach – Taking absolute temperatures from all USHCN stations, averaging them for each year for raw and adjusted series, and taking the difference for each year (the method Steven Goddard used).
2. The Common Absolute Approach – Same as the all absolute approach, but discarding any station-months where either raw and adjusted series are missing.
3. The All Gridded Anomaly Approach – Converting absolute temperatures into anomalies relative to a 1961-1990 baseline period, gridding the stations in 2.5×3.5 lat/lon grid cells, applying a land mask, averaging the anomalies for each grid cell for each month, calculating the average temperature for the whole continuous U.S. by a size-weighted average of all gridcells for each month, averaging monthly values by year, and taking the difference each year for resulting raw and adjusted series.
4. The Common Gridded Anomaly Approach – Same as the all-gridded anomaly approach but discarding any station-months where either raw and adjusted series are missing.
The results of each approach are shown in the figure below, note the spike has been reproduced using method #1 “All Absolutes”:
The latter three approaches all find fairly similar results; the third method (The All Gridded Anomaly Approach) probably best reflects the difference in “official” raw and adjusted records, as it replicates the method NCDC uses in generating the official U.S. temperatures (via anomalies and gridding) and includes the effect of infilling.
The All Absolute Approach used by Goddard gives a somewhat biased impression of what is actually happening, as using absolute temperatures when raw and adjusted series don’t have the same stations reporting each month will introduce errors due to differing station temperatures (caused by elevation and similar factors). Using anomalies avoids this issue by looking at the difference from the mean for each station, rather than the absolute temperature. This is the same reason why we use anomalies rather than absolutes in creating regional temperature records, as anomalies deal with changing station composition.
The figure shown above also incorrectly deals with data from 2014. Because it is treating the first four months of 2014 as complete data for the entire year, it gives them more weight than other months, and risks exaggerating the effect of incomplete reporting or any seasonal cycle in the adjustments. We can correct this problem by showing lagging 12-month averages rather than yearly values, as shown in the figure below. When we look at the data this way, the large spike in 2014 shown in the All Absolute Approach is much smaller.
There is still a small spike in the last few months, likely due to incomplete reporting in April 2014, but its much smaller than in the annual chart.
While Goddard’s code and plot produced a mathematically correct result, the procedure he chose (#1 The All Absolute Approach) comparing absolute raw USHCN data and absolute finalized USHCN data, was not, and it allowed non-climatic differences between the two datasets, likely caused by missing data (late reports) to create the spike artifact in the first four months of 2014 and somewhat overstated the difference between adjusted and raw temperatures by using absolute temperatures rather than anomalies.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Steve, has replied.
http://stevengoddard.wordpress.com/2014/05/10/ushcn-adjustments/
@ur momisugly Zeke
I see that now..a smaller hockey stick than Steve Goddard’s but but but..man o man that is a decade and a half of model warming manufactured right there and it only took three months Wow.
These “adjustments” falsely exacerbate warming. The past is adjusted down while the near term is adjusted up.
The bias apparent here is not in the thermometers.
The bias is with the US government agents’ conclusions about the thermometers.
Zeke Hausfather (May 10, 2014 at 8:16 am) – “Using anomalies rather than absolute temperatures isn’t adjusting the data per se. “.
It isn’t adjusting the data at all. It’s simply a way of using the data. The data consists of the original measurements, ie. the absolute temperatures. For some purposes – eg. to relate changes across different data sources – it is appropriate to use anomalies. Note that, together with their basis (which must therefore always be properly documented), anomalies still carry the original measurements, ie. the absolute temperatures. [Similarly for adjustments].
Steven Mosher – I’m a bit bemused by your (May 10, 2014 at 10:12 am) “Station moves. […] So its adjusted.“. I thought the BEST method, which you participate in, was to treat a station move (or similar) as creating a new station. In such a situation, every time an adjustment is made, a possible error is introduced. In other situations, adjustments may be legitimate, but (as for anomalies) their basis must also be kept so that the adjusted temperatures still carry the originals.
Pamela Grey (May 10, 2014 at 10:52 am) and UnfrozenCavemanMD (May 10, 2014 at 12:27 pm) and perhaps others comment along these lines too.
For those who wish to look at what BOM have done even to the GISS data in Australia.
Again, MASSIVE cooling of the past, especially the VERY WARM 1910 and 1940 periods.
As a result is very easy to set new high temperature records and have Australia’s “hottest ever”
http://kenskingdom.wordpress.com/2010/05/12/the-australian-temperature-record-part-1-queensland/
At the very least Oregon should take its stations back from any kind of national ownership and fund proper data handling. In fact every state should do that. Are we really becoming one big state with a King in power? Really? Take back our weather stations, our plains and forests, our mineral resouces, and our lands from national ownership. I for one, have had enough of this feudal set up.
Bob Greene says:
May 10, 2014 at 7:39 am
“All adjustments are negative and generally greater the farther back from present. The supposedly “problematic” 1940′s show greater adjustments in all but S. Goddard’s method. This doesn’t explain the rationale of seemingly always cooling the past with lesser adjustments as the climate “warms.””
Bob, there can be little question that there has been a lot of agenda driven adjustments. I can see why there might need to be some adjustment for station moves, etc. but the first serious adjustments made (by Hansen et al at GISS) were in 1997-8 when it became clear that the much anticipated new CONUS temperature average high WASN’T going to obliterate the pesky 1937 all time (instrumental) high. There were some emails between an underling working on the “problem” and Hansen that after a few iterations came up with the 1998 high (because of the super El Nino of that year, they realized it might be some time before there was an actual new high if they didn’t act now). I haven’t found the WUWT article on this yet but probably someone here can dig it up for you. Without this “work”, 1937 would still reign as the top average temperature today. The mid thirties to mid forties highs also showed up elsewhere, including Greenland. There has been almost continuous adjusting (as shown by the plots above) since – essentially rotating the record counterclockwise with the thumbtack ~ at 1950 or so and by pushing the “jumps” back into alignment – older ones downward, more recent ones upward – without being assured that the jumps aren’t real. Anyone with the links, please supply.
Oh. And I would take back our wildlife and the management there of. The animals that live here belong to Oregon. And the ones that are not native to Oregon will just have to go back where they came from. If we screw up, it will be our fault. I get really irritated with outsiders telling me, a life-long Oregon resident, whether or not I can use my own &^%$# forest!
Sorry, but this station thing has my Irish red hair flaming!!!!
Zeke Hausfather says:
May 10, 2014 at 1:03 pm
I am surprised that the thermometers were not regularly calibrated. Back in the 60’s when I took chemistry and physics in high school, that was one of the first things we learned. Put a thermometer in boiling water and check to see what it read (or yes, melted Napthalene is a better test). Put it in ice water and see what it read, and develop a correction formula for the thermometer it necessary. (We were also given “calibration” thermometers to calibrate our thermometers. This is like a grade 9 or 10 experiment.) I thought thermometers in weather stations were calibrated intermittently but that they didn’t realize that glass creep over a period of time could also cause issues. Did no one ever do simple calibrations by going from station to station with a calibration thermometer? I am gob smacked it they didn’t My Physics and Chemistry teachers were very clear on the need for calibration and we are talking 1959/60 so why would weather station equipment not be subject to regular calibrations? Maybe I just don’t understand as it is something I was trained to do from junior high school to high school, to university and in my engineering career.
These adjustments seem odd if there was any sort of calibration. There has to be something else like a bias of some sort or systemic error like maybe having used the wrong calibration procedure.
Many agencies require regular calibration, I am surprised this would not be done on weather stations so why the adjustment? (maybe I have not read enough of Anthony’s work) http://www.albertahealthservices.ca/EnvironmentalHealth/wf-eh-thermometer-calibration.pdf
I also am very suspicious of the Mean/Average temperatures being used after looking at many station data since the trend of the low temperatures and the trend of the high temperatures seem to paint a picture of convergence. i.e.. The stations may show flat, decreasing or slightly increasing HIGHS over 60 to 100 years+; while low temperatures show LESS COLD trends resulting in the appearance of convergence of highs and lows for the years of record. i assume this is a short term phenomenon but it does lead to a warming trend for the MEAN temperatures, Seems somewhat meaningless for most of the year : -38 C doesn’t feel a lot warmer than -40C as a low, nor does – 20 versus – 22 and the highs that stay the same at -10 or go down a degree to -11C. And summer high temperatures don’t seem too be increasing but summer lows are getting higher (in most cases). The averages give us a measure, but they don’t tell us what is happening.
But that is just my simple opinion or maybe it’s a pet peeve. But “mean” temperature is but one calculated “measurement” variable that does not tell the whole story.
Sample (note there is a labelling error on the second to last graph but the comments still apply)
https://www.dropbox.com/s/mumeba3ox98vdaj/TisdalePost.pdf
I have never seriously studied statistics, so I have no argue from authority to offer you, but I have realised that this post is just wrong on a number of counts.
Steve Goddard is looking at two data sets. One is the observed or raw data and one is the published or adjusted data. Both data sets are snapshots as far as Steve’s analysis goes as new values will be added as time passes. Therefore Steve’s approach is valid as it is comparing one data set to another at the one point in time. It is not doing a grid or station by station analysis.
The fact that stations are missing from either data set is part of the adjustments that have been done. If some stations are “late” for the raw data set, but are being given a value by the reported data set, then that is part of the adjustment. If these data points become available at a later point in time, then doing the analysis again with a different snapshot is valid, but it does not make the analysis that was done at that point in time any less valid.
Using anomalies or other data torture methods is a different analysis. Doing anything else to the data is just adding an unnecessary layer of manipulation that is not required and not really valid when you are looking at the data set to data set approach that Steve has done.
Another failed gatekeeping exercise by WUWT.
Bad Andrew says:
May 10, 2014 at 6:59 am
The fact that they adjust data at all means they aren’t measuring it correctly.
========================================================
The fact that they adjust the data is the fact that they adjust the data, regardless of measurement accuracy.
Gary, do you mean this?
http://stevengoddard.files.wordpress.com/2010/10/1998changesannotated.gif?w=500&h=355
Why are the winter month adjustments extra weird?
http://sunshinehours.wordpress.com/2014/05/10/ushcn-2-5-adjustments-final-raw/
Gary Pearse says: May 10, 2014 at 3:23 pm
“Without this “work”, 1937 would still reign as the top average temperature today.”
No, 2012 is hottest on any reckoning.
Nick,
Not really, if you look at the highly inflated bs temp records you would see that 2010 reigns supreme. 2012 is number 9 apparently and if you used raw data 1937 would be unbeatable until the next ENSO step change. I would point you at the NCDC’s climate page but its been wiped out and replaced by http://www.Climate.gov a total propaganda piece.
v/r,
David Riser
“The fact that they adjust the data is the fact that they adjust the data, regardless of measurement accuracy.”
In science, you don’t adjust data you think is accurate. To adjust data that is accurate would be a sign of mental illness. Is that what you are suggesting is happening here?
Andrew
Just a short thought regarding the Mercury-in-glass thermometers. When I started
working for the “Weather Bureau” in the early 1960’s, we used the standard USWB
thermometers, which were provided, for years by Wexler (sp). These thermometers
were etched with a serial number, and came with an individual corrections card,
generally at 5 degree intervals, valid to within 0.1 degree F. However, the Max
and Min thermometers had only a plus/minus accuracy of 1 degree F.
Gradually, due to cost considerations, the regular thermometers were allowed to
be replaced by thermometers which had an overall accuracy of plus/minus 1
degree F, as they were broken and replaced. So, as time went on, the
liquid-in-glass thermometer readings were degraded.
At larger airports, and then gradually at smaller airports, the remote thermometers
were installed (HO-60, HO-63, and in the military, the TMQ-11 series). At least, in
the early 1960s when I was working as an observer trainee, the electronics
technician went out to the instrument and used a mercury-in-glass thermometer
held up to the intake of the electronic sensor, and would compare its reading
with the temperature indicated on the readout inside the USWB office. If it was
outside the plus/minus 1 degree F range, it was adjusted back within tolerance.
I went into research for about 20 years, and came back to observing in the late
1970s. At the sites where I worked, the HO-60 series temperature sensors had
been replaced by the HO-83 series sensors due to costs, and better relative
humidity/dewpoint sensors, and the data quality checks had
been reduced to around once a month for comparison. Note that the
temperature sensor’s claimed accuracy was plus/minus 1 degree C (1.8 degrees
F). There have been many discussions concerning temperature inaccuracies
with this sensor in the literature due to deficient ventilation.
When ASOS replaced the HO-83 sensor, only a little improvement was made,
with the same problems, although they reversed the flow from in at the top…
out at the bottom in the HO-83 to in at the bottom…out at the top in the
ASOS system with a much greater strength air flow. However, the same sensor
package was used, with the same dew-point mirror with thermocooler for
dewpoint. With the increased ventilation, the claimed accuracy was stated
to be plus/minus 0.5 degree C (plus/minus 0.9 degree F. Later, a modification
to the housing was made, where a “skirt” helped to cut down the air “recycling”
around the sensor. during very light winds. This modification was redesignated
the “1088” sensor package from the HO-88 sensor package used previously.
Althougha new relative humidity sensor has replaced the cooled mirror to
compute dewpoint, as of November 2011, the power was still being applied to
the thermoelectric cooled dew point sensor.
You should realize that the temperature reading on the 1088 series sensor
is a 5 minute average, with the “center” reading, if not rejected as out of
bounds (within 2 degrees F of the average of the two preceeding and two
last readings). This center point is used as the temperature observation time
on all ASOS observations from airport sites using this instrument package.
Sorry to be so long, but it gives you some more insight to changes in the
thermometer readings in the aviation sensor department of NWS/FAA
airport sites. These sensors are used to report the highs and lows at sites
located at airports across the United States.
David Riser says: May 10, 2014 at 6:15 pm
“Nick,
Not really, if you look at the highly inflated bs temp records you would see that 2010 reigns supreme. 2012 is number 9 apparently and if you used raw data 1937 would be unbeatable until the next ENSO step change.”
Gary was talking about CONUS. I think some global numbers are coming in here. Here is a news story about 2012. Not #9.
Steven Mosher (May 10, 2014 at 10:12 am) “Lets take two types of adjustments. Station moves. A station is located at 1000m asl. It moves to sea level. This will create a false Warming signal. So its adjusted.”
Earlier Zeke said: “station moves are a mixed bag, but in the 1940s and earlier many stations were moved from building rooftops in city centers to airports or wastewater treatment plants, also creating a cooling bias.”
Stations on rooftops are not necessarily warmer. They usually were placed on tall platforms well above the roof and would mainly be warmer on radiational cooling nights. The move may or may not show up as a discontinuity in the difference of means from the site to the region. But it probably doesn’t matter much.
After the move to the airport we would generally see warming both in microsite and sometimes in the vicinity. In the case of Reagan National (not in USHCN, thank goodness for small favors) they keep adding gravel around the ASOS sensor that has increased year after year: http://shpud.com/weather/main.php?g2_itemId=155 and http://shpud.com/weather/main.php?g2_itemId=158 slowly raising the average presumably without discontinuities.
The observable effect is that with calm winds (no influence from the nearby Potomac) the temperature will actually rise on radiational cooling nights when the wind goes calm (whereas it will modestly drop with continuing winds). That same effect will generally not occur at Patuxent and Annapolis (both also on the water) or any other area stations. Only Baltimore’s Science Center has a worse-sited thermometer in the DC area.
If there are groundskeepers at National Airport, the “station of record” for DC slowly warming the site, they can certainly exist elsewhere.
Other problems are increasing parking areas at airports. At national new parking areas are relatively close by since it is a cramped location. Other airports have increasing urbanization following 70’s and 80’s exoduses to the suburbs.
There does not seem to be a good solution for dealing with this data prior to 1979. But for most purposes data after 1979 should be obtained from the satellite record, for areas like the continental US and larger.
Scientific disagreement is not fraud, but in the poisoned waters (thank you Michael Mann and Co.) of climate debate, paranoid ideation occurs spontaneously when, virtually without exception, all the old temperatures get downward adjustments, while all the new temperatures get adjusted upwards. The vigorous inquiry in this thread is refreshing and makes me question my own reflexive skepticism and look harder at the actual science. More would be accomplished if others would follow suite, and shut up and calculate, as no one looks dumber than when they are cheer-leading phenomenology.
The cooling of the past seems to be systematically applied to a large proportion of the minimum temperatures.
http://landscapesandcycles.net/image/75157969_scaled_419x316.png
http://landscapesandcycles.net/image/75158304_scaled_576x434.jpg
Unhomogenized quality controlled data for the USA shows cyclical temperatures as one would expect from cycles of the Pacific Decadal Oscillation.
http://landscapesandcycles.net/image/75159090_scaled_578x580.png
http://landscapesandcycles.net/image/75158484_scaled_590x444.png
So, let’s see: Every time we come out with a new upgrade of the world’s historical climate database, 1936 get adjusted down another couple of tenths of a degree. Imagine, historical record temperatures are just like money – they get blown away by inflation. The 1930s now look sort of like the Little Ice Age; someone run and check – was London, in 1815, under a mile of ice?
The USHCN needs to be sent to prison, along with just about every other person who got named in the ClimateGate emails, and maybe a few since then. They are crooks, charlatans, liars, communists, bad scientists, or some combination of all those things. 20 years at a minimum. And I mean it.
“You really should issue a correction on your blog to make the reason for the spike clear, that it isn’t “tampering” per se, but an artifact of missing data and your method – Anthony”
“Yes, but it’s wrong, so learn from the mistake, issue a correction and move on. – Anthony”
So, what do you think of Goddards “correction” Anthony?
So here we have the Bullshite reason for adjusting past temperatures lower “Back in the 1940s virtually all the stations used liquid-in-glass thermometers, which read about 0.6 degrees warmer in max temperatures (and about 0.2 degrees colder in min temperatures) than the new MMTS instruments introduced in the 1980s.”
Now I have worked in “Quality” all of my working life, including Metrology labs, under no circumstances would it be acceptable to go back and change a hundred years of “Calibrated” values Ad Hoc because a new measuring method was obviously WRONG. The MMTS thermometers should have been calibrated to the Glass/Mercury ones, not used to thorw a hundred years of readings in the bin.
Does Zeke actually believe that the Victorians were so bad at Engineering & Science that they could not get a Thermometer scale right, it just beggars belief.
As others have mentioned Glass/Mercury Thermometers were accurate and I would far rather believe one than an “Electronic” device, anyone tried taking their temperature with a modern Thermometer to see what crap some of them are?
James Hall (NM) says:
May 10, 2014 at 6:32 pm
says it all for me, he speaks from experience, add to that all the siting problems we know about and “natural” differences in temperatures within a few hundred yards let alone Miles you can see that each Thermometer measures a REAL temperature for where it is.
Just because it doesn’t agree with another thermometer 10k away does not make it wrong or in need of “Correction”.
Nick Stokes says:
May 10, 2014 at 2:05 pm
It’s just wrong method.
No it is exactly the right method to show the “Manipulation” that is taking place and I am not just talking about the so called Corrections, I am talking about the presentation of the data to the Public to push an Agenda. You can deny it as much as you like but most people on here do not believe you.
It is quite odd how the same people that have argued that UHI does not affect the Global Temp trend but arre now saying adding in shorter length records or thermometers with broken records does.
Perhaps they are prepared to present their findings for this?
WUWT study shows that “Poor Quality” sites are a bigger problem and 1000km “Gridding” is complete joke.
The BEST minds in climatology do not know what an absolute temperature ist?!