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.
[snip – if you have criticism, explain it. I’m not going to tolerate any more of your drive by crypto-comments that contain the word “hint”. For a person who’s always on about full disclosure with data and code, you sure do a crappy job of following your own advice when you make comments – Anthony]
Zeke Hausfather says:
May 10, 2014 at 8:16 am
Using anomalies rather than absolute temperatures isn’t adjusting the data per se. Homogenization does adjust the data, but the alternative is only using stations with no moves, instrument changes, time of observation changes, etc.
Individuals have analysed some of those adjustments and shown that they are incorrect when comparing to other local stations, but no one goes back in to the system and removes or corrects the adjustment process. The mass adjusment process is crap, as has been shown many times, one of the worst cases being Iceland, which are adjusted after already being adjausted by their process.
Also the classification of Rural/Urban etc is completely broken as has been proved many times.
The notion of an “Average Global Temperature” is nonsense and when handled by people with an agenda can and has been shown to be manipulated to prove AGW.
If the world needs to know whether it is warming or not you can take each station, calculate any change over a period, noting any “Changes” made to the station and then let people make up thier own minds, not force a gridded, homogenized and badly adjusted version on us.
“Its good for clarity, it doesn’t change the message.”
This is what he actually said.
“Two things stand out about the current USHCN data tampering graph. The most obvious is the huge amount of tampering going on in 2014, but almost as bizarre is the exponential increase in tampering since about the year 1998.
*insert graph that purports to show the exponential increase in tampering*
There is no rational reason for either of these – so here is my guess. Obama wants credit for healing the climate. He has been engaging in every imaginable form of BS to get an international agreement through this year or next, and after he gets the agreement he will tell NOAA to stop tampering – and will then take credit for the drop in temperature.”
Just being technically correct in a certain way with the graph doesn’t excuse his attempt to misrepresent what it means.
“The notion of an “Average Global Temperature” is nonsense and when handled by people with an agenda can and has been shown to be manipulated to prove AGW.”
You do realize that you are talking to one of the people who actively works on that subject?
Anthony,
I agree with your assessment of what causes the spike mathematically, and I saw the same thing. I don’t agree that it is proper for USHCN to fabricate adjusted data for stations where they don’t have raw data.
REPLY: yes, which speaks to the numbers the release in the monthly State of the Climate Reports, which are released with US historical temperature rankings for that month before all the data is in. Invariably, those rankings change later.
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
drumphil says:
May 10, 2014 at 9:11 am
You do realize that you are talking to one of the people who actively works on that subject?
Yes and do you think he has adequately explained the 1 degree cooling of the past in his various expanations?
tgasloli said at 8:42 am
There is no other field in science where the data is routinely corrected in one direction only. …
B I N G O !
And it’s not just temperature. Ocean Heat and Sea Level are also adjusted and I’m sure evry other aspect of Global Warming from polar bears to to hurricanes and glaciers is similarly fudged. Winston Smith and the Animal Farm pigs have nothing on these guys.
Zeke Hausfather says:
May 10, 2014 at 8:16 am
David Riser,
Using anomalies rather than absolute temperatures isn’t adjusting the data per se.
>>>>>>>>>>>>>>>>.
No it isn’t. But both methods mask the real problem. We’re trying to understand how CO2 affects energy balance. As energy flux varies with T raised to the 4th power, averaging either absolute values or anomalies simply dispenses with this fact and produces a trend that is at best only loosely related to the problem at hand. In fact, there are circumstances when the trends in average energy flux and average temperature (or anomalies of temperature) can be in opposite directions.
Unless and until the laws of physics are applied to the data, all your work to produce an accurate temperature record will result in little more than a more accurate trend of the wrong metric.
The other thing that I question about “Global warming” is how can warming be “global” when it is not even “Continental”, let alone hemispherical?
How many times have we seen one half of the USA or Europe warm while the other half cools. To me global means everywhere at once.
David Riser,
Goddard’s problem is that he is just averaging all the absolute temperatures together each year. Everyone else (Anthony included in his papers) use anomalies, because absolutes run into lots of problems when the set of stations is not consistent over time. That why all three other methods we explore (absolute temps with forced consistency, anomaly methods) all get pretty much the same result, and why they are all fairly different from Goddard’s result.
R2Dtoo,
We all know the raw data is biased because of station moves, instrument changes, TOBs changes, etc. The question is can we effectively detect and remove these biases using statistical techniques without adding any additional bias in the process. The method used by NCDC (and Berkeley) is conceptually simple: look at the difference between a station and all its surrounding neighbors, look for step change breakpoints at one station that are not shared by the neighbor, and flag that as an inhomogenity. For example, if one station changes its instrument from liquid-in-glass to MMTS in January 1984, but most of its neighbors don’t make the same change at the same time, this pairwise approach will detect a step change at that station in 1984 and remove it (or cut the record and treat everything afterwards as a separate station in the Berkeley approach). This will generally work unless all the surrounding stations change at the same time, something that is fairly unlikely given that the major changes (TOBs, MMTS) were all slowly phased in over 5-10 years.
For folks saying that dealing with biases in raw data is somehow unique to climatology, I’ve had to deal with it in other fields as well. For example, databases of manual electricity meter reads contain all kinds of weird artifacts, missing data, or errors that need to be corrected during data normalization that, if not addressed, would skew the results of a disaggregation analysis. What is needed is a consistent, automated, well-documented, and tested approach to dealing with errors in raw input data. That way if you discover that your algorithm isn’t working properly, its easy enough to make changes and rerun the whole dataset. You start running into problems when you make one-off or manual adjustments to individual stations, because you don’t necessarily end up with consistent standards.
I have noticed NCDC announcing “nth warmest” statements about the previous month. In all cases the satellite data does not seem to agree with these statements. Are these statements also an artifact of missing stations? If so, I would have to say the Goddard has a point. NCDC is making false statements about the climate. Since that is the result of their own process, it is by definition .. “tampering”.
Richard M,
That is correct. They will usually go back and update it once the data is in but there isn’t any discussion or announcement. Its one reason the top 10 appear so often. NCDC publishes it as a top 5 or 10 then changes is to where its supposed to setting up the next opportunity for a top etc. Its one reason I have issues with adjustments. Steven Goddard’s work demonstrates this neatly and yes the spike looks bizarre but it doesn’t stop NCDC from doing what they do.
So Zeke I get your point but you don’t get mine, so we just have to agree to disagree.
v/r,
David Riser
David Riser says:
May 10, 2014 at 9:53 am
I notice that Zeke does a lot of “justifying” the adjustments, but has not really commented on the at least 1 degree cooling of the past that they appear to have caused, as most have noted error adjustments should never all go in one direction, unless by mistake or by design.
Ima.
The reasons for adjustments are real.
The adjustments are validated
The adjustments have been investigated by skeptics and vindicated.
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.
Next a ststion changes instruments. Side by side
Tests are run an adjustment is created.
Lastly the station changes tob. This bias is removed
By a tobs change.
Tobs adjustments are the biggest.
Those adjustments were verified many times including
At john daly site.
REPLY: Sure, that’s an opinion, but the stations that don’t need adjustments to fix problems is where the true signal exists, everything else is just guesswork and “best estimates” with error bars. Adding more noisy stations to the mix does not improve the accuracy of getting the sought after AGW signal, it only increases the uncertainty.
You’ll soon be having to revise your world view. In the meantime, please limit your drive-by expositions here unless they contain something of substance, backed up by some citation. – Anthony
Is there a published reason why past temperatures have been adjusted downward? Or is there even an acknowledgement by the climate science community that this is happening?
So is this statuon data part of the data used for this analysis?
http://www.foxnews.com/science/2013/08/13/weather-station-closures-flaws-in-temperature-record/
NCDC reports absolute temperatures, not anomalies. Most of my comparisons are vs. GHCN HCN vs. NCDC. Once in a while I do the USHCN comparisons like this one.
REPLY: Yes, but it’s wrong, so learn from the mistake, issue a correction and move on. – Anthony
It is interesting that the adjustment in 1940’s US temperature data is greater than the difference between the MWP and the LIA in the famous hockey stick graph. So the measurement error in 1940’s US temperature data is claimed to be greater than Michael Mann’s estimate of the actual range of Northern Hemisphere average temperatures over almost 900 years. Isn’t amazing how stable Northern Hemisphere average temperature was for the 9 centuries preceding the introduction of significant CO2 by humans? Or, rather, isn’t it amazing how gullible the IPCC, the press, and public were when the hockey stick graph came out.
A botched plot never roils.
Mr. Mosher,
I would say adjustments have not been vindicated. I would suggest the opposite is true. The difficulty with most of what you say is; the adjustments are automated within a grid, they are not found by reviewing the thousands of station records and doing a station by station adjustment. The grid resolution is too large to effectively capture all the changes that can occur within that grid, the fact that 1000 ft. differential in altitude can and does exist within your grid demonstrates the fallacy of which you speak. Tobs is a tiny issue that you all assume was unknown in the past yet I am pretty sure the folks doing the observation were very aware of what they are doing, so finding tobs changes in the record by looking for step changes in the data and doing a massive rewrite of that data is not an acceptable practice.
v/r,
David Riser
Mosher, I don’t think they should be adjusted at all. There should instead be a break in the data indicating changed conditions (IE construction of a road over what was once grass, or a new building goes up, or a new kind of unit). And when a station is moved from its pole stuck in the ground -up, down, or sideways- it should be declared dead. End of data for that station. I think a new data series should be put together with these criteria, and a new set of station ID numbers provided per the above criteria. Any other research entity (IE agricultural seed plots) gathering subject data would do this. That climate researchers don’t is just plain laziness or a desire to create long records where none exist. I am thoroughly unimpressed.
Pamela Gray says:
May 10, 2014 at 10:52 am
The unwarranted adjustments are made for the express purpose of lying. Same goes for the notations that should be made along your suggested lines but aren’t. I hope that enough original observational data, not adjusted, bent, folded, spindled & mutilated by mendacious, trough-feeding climastrologists, remain intact for a more honest temperature record someday to be constructed for purposes of actual science rather than warmunista advocacy.
We should note also that from the website:
http://www.ncdc.noaa.gov/oa/climate/research/ushcn/
we discover that the data is not adjusted it is “corrected” (following is a direct quote):
“USHCN temperature records have been “corrected” to account for various historical changes in station location, instrumentation, and observing practice.”
The word “corrected” is in quotes. I guess they couldn’t figure out how to get a “wink” emoticon to display.
🙂
NikFromNYC
You say
thus many conservative bloggers too often pick up on his posts
Do you mean too = to = also
or too as in too many = an excess?
It’s not clear from your text and to and too and two are in fact three different words with different meanings.
Any chance that a couple of these records be maintained here on WUWT? Suggest that USHCN data be used, gridded, with missing data not counted, no other adjustments. It should be stand-alone, without reference to USHCN computed data which changes whenever somebody gets the idea for a new “validated” adjustment. After a certain interval it could be considered “research grade.” The simple raw average should also be maintained.