Guest post by David Middleton
Featured image borrowed from here.
If you can set aside the smug, snide remarks of the author, this article does a fairly good job in explaining why the surface station temperature data have to be adjusted and homogenized.
There is just one huge problem…
Zeke Hausfather/Berkeley Earth”… I added the the natural variability box and annotation. All of the anomalous warming ince 1960 is the result of the adjustments.
Without the adjustments and homogenization, the post-1960 US temperatures would be indistinguishable from the early 20th century.
I’m not saying that I know the adjustments are wrong; however anytime that an anomaly is entirely due to data adjustments, it raises a red flag with me. In my line of work, oil & gas exploration, we often have to homogenize seismic surveys which were shot and processed with different parameters. This was particularly true in the “good old days” before 3d became the norm. The mistie corrections could often be substantial. However, if someone came to me with a prospect and the height of the structural closure wasn’t substantially larger than the mistie corrections used to “close the loop,” I would pass on that prospect.
Just for grins, I plotted the UAH and RSS satellite time series on top of the Hausfather graph…
I think I can see why the so-called consensus has become so obsessed recently with destroying the credibility of the satellite data.
Addendum
In light of some of the comments, particularly those from Zeke Hausfather, I downloaded the UAH v5.6 “USA48” temperature anomaly series and plotted it on Zeke’s graph of US raw, TOBs-adjusted and fully homogenized temperatures. I shifted the UAH series up by about 0.6 °C to account for the different reference periods (datum differences)…
I used a centered 61-month average for a 5-yr running average. Since there appears to be a time shift, I also shifted the UAH ahead a few months to match the peaks and troughs…
The UAH USA48 data do barely exceed the pre-1960 natural variability box and track close to the TOBs-adjusted temperatures, but well below the fully homogenized temperatures.
Your figure with RSS and UAH data look quite odd; are those their U.S. lower 48-only numbers?
UAH doesn’t seem to have U.S.-only processed data available for their new beta version; for the current operational version (5.6) they do, and they are available here: http://vortex.nsstc.uah.edu/data/msu/t2lt/uahncdc_lt_5.6.txt
If I plot UAH 5-year averages on my graph (which gives me from 1983-present, since prior to 1983 there isn’t 5 years of UAH to average, as it begins in 1979) for all the series from 1983 to present, aligning all series to the 1983 5-year average, I get this. UAH at least (version 5.6) seems to agree much more with adjusted data than raw data, and looks nothing like you graph.
http://s24.postimg.org/pb575mexx/tavg_ushcn_raw_tobs_adj_v3.png
If you know where I might find U.S. data for UAH v6, I’d be happy to plot that as well.
Also, you neglected to show the figure featuring global adjustments. Turns out you actually get less warming in the adjusted data, not more. Funny that.
http://cdn.arstechnica.net/wp-content/uploads/2016/01/noaa_world_rawadj_annual.png
Thanks to USCRN, all US temperatures have been brought under some semblance of adjustment control.
Trouble is, that leaves the rest of the world for the alarmista to play with. And they do.
The cited article (at ArsTechnica) begins, as its very first example of the scientific need for adjustments, the measurement of groundwater level over time. This example contradicts the article’s argument, stating: “Automatic measurements are frequently collected using a pressure sensor suspended below the water level. Because the sensor feels changes in atmospheric pressure as well as water level, a second device near the top of the well just measures atmospheric pressure so daily weather changes can be subtracted out“.
Thus the device collects two raw measurements, subtracting one from the other, producing a “raw” difference between the two sensors. This is the design of the apparatus. Neither sensor is adjusted.
David Middleton closes with “I think I can see why the so-called consensus has become so obsessed recently with destroying the credibility of the satellite data. Indeed. The so-called consensus has claimed that satellite also “requires adjustments,” ignoring the fact that — unlike ground weather stations — the “adjustments” are part of the measurement design, not after-the-fact fiddling.
Thank you for a very informative article, and also for linking to the ars technica article, which as you say has a snarky tone but does a good job of broadly describing the whys and wherefores of homogenisation.
You should read the comments forums to this article.
I comment frequently in their forums when they are on climate topics which they often are.
They are a fully committed to the cause crowd, enforcing the AGW dogma.
If one offers any dissenting ideas they are immediately attacked, some times correctly so. But in other cases cogent arguments are made, then there is a coordinated squad of enforcers that will attack, eventually devolving to swearing at the poster with assorted an hominums.
You should look at the comments to that story and see the partisan tatics.
Since when was the argument that temperature record was good as is? A quick look at nearby stations shows how you couldn’t be sure of anything unless the change was tenfold bigger and you restricted yourself with rural sites with few changes.
The argument is that with such large adjustments that you can’t have much confidence in a result that shows such a small trend globally and far from uniform across the globe. Even without accusations of fudging, it still isn’t good enough for basing policies on.
But the result of the adjustments is not the data being offset across the range, a small increase in the trend over the whole range nor is the plot of the differences from previous estimations noisy. The difference is a very smooth plot of what the activists wanted to see.
Back in the day, we would call NOAA “Shits and Giggles.”
Ha ha
Even today they can’t figure out the meaning of an arithmetic mean from a mean.
Don’t ask a NOAA employee what is a geometric mean to an arithmetic mean! That, at a bar in Bethesda, would start a fight and the police would be called to close the bar.
Ha ha
News Flash: Climate change caused by newly discovered super planet with orbit of 20,000 years. The planets are aligned right now. Big changes expected. Lost Pluto but found Micky Mouse.
Problem with USHCN : The logic around TOBS adjustment appear sound, but the documentation that observers actually did change their time of observation as needed for the huge adjustments from 1970, to 1980 to 1990 to 2000 to 2015 appear weak. So the foundation is weak.
When at the same time the number of missing datapoint in the most recent years has exploded, it becomes increasingly hard to explain adjustment as usual: “Well in the old days, collecting of temperature data was so very bad..” style.
On top, it seems that Karls UHI writing 1986 for USA has been forgotten and people are forced to believe this issue is tiny.
I understand homogenisation to get a global temperature. I understand adjusting data to get a consistent set of data for comparisons. What I don’t understand is where the “better” data comes from that is used for homogenisation and adjustment.
If we have lots of high quality data showing this warming, lets see it. If we don’t then you can’t use lower quality data to adjust higher quality data and claim you end up with a better result.
I don’t care what field you are talking about, you just cannot.
Why is this post only discussing the effect adjustments and such have on temperatures for the United States when one gets radically different results if they look at temperatures for the entire planet? Did the author just not look at what happened with global results, or is this just massive cherry-picking? Either way, it’s pretty ridiculous. Whatever value there may be in looking at US temperatures, there would be far more value in looking at global ones.
Especially since the post compares the temperatures it examines to global satellite temperatures. Why use global temperatures in one case but US temperatures for the other? Given basically nothing the post says holds true for global temperatures, it just looks like massive and intentional cherry-picking to me.
I think I can see why the so-called consensus has become so obsessed recently with destroying the credibility of the satellite data.
It’s the only thing keeping them remotely honest. Without the sanity check from Satellites they have nothing holding them back.
Again we forget to show that FOUR radiosonde datasets agree with the TWO satellite sets. There is no argument and NO warming
http://www.globalwarming.org/wp-content/uploads/2013/06/CMIP5-73-models-vs-obs-20N-20S-MT-5-yr-means1.png
Re: Thoroughly fabricated … data, 1/21/16:
[W]hy the surface station temperature data have to be adjusted and homogenized.
Data evolve to fit the model, to earn tenure, to make the catastrophe really scary, to loosen the money, to pay the salaries, to buy the next gen computers, to regulate out the capitalists, to elect the socialists.
It’s not wrong — it’s Post Modern Science.
Here is where it all began:
With 1998 El Nino in, it was clear that 1934 in the US was still the record high year. Knowing that this was likely to be followed by a La Nina, it was likely going to be a long time they would have to suffer with the embarrassing fact that 1934 remained the record. The email from an FOI request shows chronologically how his assistant adjusted the figures until Hansen was happy!! I urge everyone who is wondering about this issue to read the historical first egregious tampering with the official record.
Remember that Hansen in 1988 had the airconditioning turned off and the windows all closed the night before in preparation for his alarmist speech to a sweltering Congress (an obliging congressman [Wirth??], I believe arranged this). Here, ten years later with the 1934 warmth dogging hime, he showed another example of his lack of scruples. Today, we have largely forgotten this historic fact of tampering with official US temperatures. We agonize over and measure the comparatively small changes made by T. Karl and whether the pause is 18 years or whatever, when the pause might in actuality be 80 years. Everyone argues that the US is only 3% of the Globe so it doesn’t mean that globally it is similar. However, the Iceland, Greenland, Canadian and Northern Russian temperature records also had 1930s-1940 as the warmest. WUWT? The Canadian all time high was in two places in Saskatchewan in 1937 when it was 47C!! and it was in the high 30s and 40s throughout much of the rest of Southern Canada as well.
thanks, when I dropped that into tip line yesterday
http://wattsupwiththat.com/tips-notes-2/#comment-2125475
I hoped someone would address it
It seems curious to me that the adjustments made post 1980 are significantly larger than the adjustments made pre-1980. Did we just not know what we were doing 1980 – present to cause the temps to be consistently measured lower (and a lot lower at that)?
I live in that area of Saskatchewan that had 47C in 1937 although I’m not that old. The local weather regularly shows record highs for a particular date as occurring in the 30’s. As well, we had significantly hot, dry weather in the 1980’s and I certainly remember some serious heat in the 60’s when I was a kid. Outside all the time and no A/C in most buildings might affect my memories but I’m curious about the 16-17 year gap between major el nino’s. Does this interval show up in records previous to 1998?
It’s not a “claim.” It is a fact. Identifying the problematic portion of an article is not cherry-picking.
Your point about my not going way overboard in clearly identifying where I was being snarky is valid. And clearly I should have used much larger, eye-catching fonts when I clearly stated that I was comparing the global satellite data to the US surface data.
I have since appended the post with a comparison of the UAH satellite data for the US and Zeke’s graph from the article. The result still illuminates the so-called consensus’ obsession with destroying the credibility of the satellite data.
Curious George
January 22, 2016 at 9:21 am
“Could you please provide a link?” [Query re comment above on 1934 still hotter than 1998 according to Hansen:
http://wattsupwiththat.com/2016/01/21/thorough-not-thoroughly-fabricated-the-truth-about-global-temperature-data-well-not-thoroughly-fabricated/#comment-2126590%5D
Here is the original link to an expose from an FOI request concerning the fiddling with 1934 record hot temperatures to reduce them below that of 1998 and the anger and excitement it garnered. Sorry I forgot to add it to my comment.
http://wattsupwiththat.com/2010/01/14/foiad-emails-from-hansen-and-giss-staffers-show-disagreement-over-1998-1934-u-s-temperature-ranking/
We have to keep hammering this stuff!!
Gary Pearse wrote:
“With 1998 El Nino in, it was clear that 1934 in the US was still the record high year. Knowing that this was likely to be followed by a La Nina, it was likely going to be a long time they would have to suffer with the embarrassing fact that 1934 remained the record.”
Can someone please tell me why El Nino increases global mean temperature and La Nina decreases it? according to the model they do not change mean temperature, they just redistribute it around the globe
Deal to this guys. It is important. Why? because should a logical explanation not be found it proves that surface measurements are not accurate. Do the satellite measurements show the same correlation? If they don’t then ‘bingo!” – you’v found the strongest arguments yet that surface measurement are biased by environment. How the heck could 18 months of El Nino increase mean marine temperature by any degree measurable??
Another option: not thoroughly enough fabricated …
Dr. Roy Spencer posted radiosonde data in Figure 7 of
http://www.drroyspencer.com/2015/04/version-6-0-of-the-uah-temperature-dataset-released-new-lt-trend-0-11-cdecade
This indicates the surface-adjacent lowest troposphere warmed since 1979 by about .02, maybe .03 degree C per decade more than the satellite-measured lower troposphere.