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.
Nick Stokes says:
May 11, 2014 at 9:55 pm
“No, you can’t tell that way. Suppose a warm Monday afternoon, peaks at 4pm. The 5pm reading will say that 4pm is max for Monday, but the 5pm will be max for Tuesday.”
OK I accept that “could” happen.
So are you going to correct for it, you say we must?
But it didn’t happen and it didn’t happen 6 days that week and it certainly didn’t happen for months in the winter.
So you have now “corrected” 6 days that week and months in the winter that didn’t need it.
The reason for our differences is you believe you know what was happening 80 years ago and we say you don’t know what the actual weather was doing for any given day in any given location.
The next question is do you hourly readings within a kilometre of those Co-Op weather stations that are read at 5pm to verify what you are sayung is correct and what I am saying can’t be?
Some years ago I was in a marine lab where the instrument recording outside temperature was electronically recording in real time. It was fairly easy to observe max and min daily temp and the mean temp was also automatically deduced. I asked why we still relied on (Tmin + Tmax)/2 as a proxy for Tmean and Steven Mosher on this site put me right.
At the time I did not realise that the Tmaxes (and Tmins) were still being recorded only at specific times of the day. The thermometer which we had at school recorded Tmax and Tmin for the whole previous 24 hours (requiring it only to be read at approximately the same time each morning). Having read this thread I am more confused than ever. Why is it necessary still to use instruments for the measurement for temperature which require TOB? Jobs for the boys?
The system has changed over to contiuous readings now, but the TOBs is used to “Correct” the old Max/Min once per day readings.
My problem with that is they don’t know if a correction was necessary or not, but they seem to apply it anyway, I am still trying to get my head around Nick’s posts on it on his Forum.
Goddard’s analysis and conclusion are entirely correct – since 1890 temps have declined on average, yet the US gov’t still pushes the hockey stick as ‘science’. Fraud, data tampering, mendacity are not science. You can quibble over esoterics – is the inflammation a lot or just enough to support more government regulation over a carbon economy? This is splitting hairs. Instead of arguing in public why didn’t WUWT simply communicate with Goddard in private, post the results and say the discrepancy might be due to xyz [I don’t think we know] ? Why would you try to devalue Goddard’s work which is correct and relevant ?
Ferdinand (@StFerdinandIII) says:
May 12, 2014 at 8:02 am
This is splitting hairs. Instead of arguing in public why didn’t WUWT simply communicate with Goddard in private, post the results and say the discrepancy might be due to xyz [I don’t think we know] ? Why would you try to devalue Goddard’s work which is correct and relevant ?
It is neither correct nor relevant! As pointed out above it is an improper averaging of different sets of data, if the analysis is done properly with the same sets of stations the large spike disappears.
take out 2014 and you still have proof that there is bias in the adjustments …
also since nobody knows what the late reporting stations are in 2014 isn’t it just a guess that the spike is caused by late reports ?
#1 is the simple, common-sense approach. It answers the question: “what is the actual effect of homogenization and other adjustments”?
#2-4 all share the same problem: they cannot answer the above question. News flash for Zeke: no one believes you guys are doing any of the adjustments honestly, even the station dropouts. Sorry, but your fellow climate scientists have dug a deep credibility hole.
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
This is true, but so what? Everyone knows the year is incomplete. Yes, the line will probably spike less by the end of the year, but it’s obviously a YTD value.
We know when the coop stations recorded their data from their records! So despite the directives we know the coop stations didn’t follow them, and that the TOBS changed over time, the effect of it was analyzed and a correction formula determined from that analysis so that the changes in practice over time could be corrected for.
And the correction formula just keeps getting warmer and warmer. If you only look for your keys under the lamppost, there’s a 100% chance you will never find them anywhere else.
The tobs adjustment really is pure bs. It doesn’t make a hill of beans if the person recording the min/max did it at a earlier time than midnight. The only thing that matters is, did they record the temp after the high for the day. If any of you had spent any time doing daily temp records you would know that for any given area the max min is very close to the same time of day. So if my max peaked at 4pm and my min is at 6am, as long as I reset the thermometer before I go to bed its all good. I am sure that the coop volunteers did exactly that. So before you go blabbing about what if is, why don’t you make some daily readings, say every hour or so and see what you actually get. what you will find is that TOBS is an excuse to freeze the world after the fact.
A C Osborn says: May 12, 2014 at 6:05 am
“it certainly didn’t happen for months in the winter:
No, it can easily happen in winter. Here’s one way to think about it. An extreme case. Suppose the max always happens at 4pm, and you read and reset at 4pm every day. Then when you write down the max, it will be the higher of now or 4pm yesterday. Best of two. Summer or winter. Higher values count for two days. Minima only once. You were “warming the past”.
If you read at 6am, and that’s when minima occurred, you’d get a cool bias. And if you change from 4pm to 6 am, that’s a big change.
TOBS isn’t applied to daily readings. It’s a correction applied to monthly averages. It would apply a different correction to different seasons, if that’s what the data says.
Nick writes “Then when you write down the max, it will be the higher of now or 4pm yesterday. Best of two. Summer or winter. Higher values count for two days. Minima only once. You were “warming the past”.”
However being realistic again, if you wrote down the max at 4pm and then it increased some more that evening and the next day was even higher then the way the TOBs adjustment works now, you’ll artificially cool the average of those two days when in fact the average should have been adjusted up.
Nick,
That’s not how a min/max works…… You take your reading late or early doesn’t matter as long as your consistent. You avoid it during min/max the thermometer does the work. So typically a coop person would visit their thermometer once a day after 5pm or perhaps later. But the typical highest temp will have been much earlier. As long as your consistent throughout the month there is no bias. The instructions have always been very simple and if they do request a TOBS change its logged and its a one time deal that if done carefully, once again does not introduce any bias. All this bs of “if this time and reset such and such” means nothing! In fact to assume that the coop observer can’t figure this out when its been discussed since 1800’s is unbelievable arrogant. To accuse them of malfeasance without evidence is criminal. To adjust their data with statistics and without ever looking at their records is stupid.
v/r,
David Riser
David Riser says: May 12, 2014 at 8:58 pm
“In fact to assume that the coop observer can’t figure this out when its been discussed since 1800’s is unbelievable arrogant. To accuse them of malfeasance without evidence is criminal”
No-one is accusing anyone of malfeasance or any kind of failure. It is assumed that the Coop observers are faithfully following the agreed procedure. The TOBS is calculated on what they reported.
There’s nothing theoretical about the TOBS effect. You can simulate a min/max reading system using hours and hours of modern data, and quantify the bias precisely. I’ve described that here.
Nick writes “There’s nothing theoretical about the TOBS effect.”
Apart from the assumption that the readings are taken at the agreed time you mean?
“It is assumed”
Climate science in a nutshell. Thanks, Nick.
Andrew
TimTheToolMan says: May 13, 2014 at 6:08 am
“Nick writes “There’s nothing theoretical about the TOBS effect.”
Apart from the assumption that the readings are taken at the agreed time you mean?”
Actually, not even that. I posted this graph of the change of reading time over the years, from Vose 2003. It shows the metadata time (what people said) and dotted curves which are “method of DeGaetano”. This is a clever analysis based on the temperatures that were reported at the time of reading. With thousands of temperatures reported, and a good handle on diurnal variation from thousands of hourly readings, you can make a very good estimate of whether reporting times are correct. And the plot shows pretty good agreement.
“clever analysis”
“good handle”
“very good estimate”
“pretty good agreement”
Nick, can I hire you as my latex salesman?
Andrew
Nick also wrote “And the plot shows pretty good agreement.”
How does that look when compared to ALL the adjusted data rather than just the 30 years you’ve chosen?
TimTheToolMan says: May 13, 2014 at 7:43 am
DeGaetano, Arthur T. “A method to infer observation time based on day-to-day temperature variations.” Journal of climate 12.12 (1999).
Nick,
If they followed the agreed TOBS there is no discrepancy or bias. I am thinking that yall need to actually observe the temperature over the course of a day for an entire year for each degree of latitude to understand where I am coming from. Climate scientists are making a mountain out of this because it conveniently fits their idea of what is. Honestly if the record is such a train wreck then yall aught to toss it out and start over, oh yea we have satellite now, so no need. hmm and the temps are flat, what do you know.
v/r,
David Riser
Nick writes “They could write down random temperatures too. But these are Coop volunteers. They go to a lot of trouble to contribute. There’s no point in doing that haphazardly.”
It seems understanding human nature isn’t your thing, Nick 😉
Straight up from DeGaetano we have…
“Introduction
Differences in observation time among U.S. climatological stations arise due to the voluntary nature of the Cooperative Observer Network. Daily maximum and minimum temperature observations from the cooperative network are usually taken at an hour that is convenient for the volunteer”
Thanks for poionting me at the reference though. That should make for interesting reading.