Dr. Roy Spencer proves what we have been saying for years, the USHCN (U.S. Historical Climatology Network) is a mess compounded by a bigger mess of adjustments.
==============================================================
USHCN Surface Temperatures, 1973-2012: Dramatic Warming Adjustments, Noisy Trends
Guest post by Dr. Roy Spencer PhD.
Since NOAA encourages the use the USHCN station network as the official U.S. climate record, I have analyzed the average [(Tmax+Tmin)/2] USHCN version 2 dataset in the same way I analyzed the CRUTem3 and International Surface Hourly (ISH) data.
The main conclusions are:
1) The linear warming trend during 1973-2012 is greatest in USHCN (+0.245 C/decade), followed by CRUTem3 (+0.198 C/decade), then my ISH population density adjusted temperatures (PDAT) as a distant third (+0.013 C/decade)
2) Virtually all of the USHCN warming since 1973 appears to be the result of adjustments NOAA has made to the data, mainly in the 1995-97 timeframe.
3) While there seems to be some residual Urban Heat Island (UHI) effect in the U.S. Midwest, and even some spurious cooling with population density in the Southwest, for all of the 1,200 USHCN stations together there is little correlation between station temperature trends and population density.
4) Despite homogeneity adjustments in the USHCN record to increase agreement between neighboring stations, USHCN trends are actually noisier than what I get using 4x per day ISH temperatures and a simple UHI correction.
The following plot shows 12-month trailing average anomalies for the three different datasets (USHCN, CRUTem3, and ISH PDAT)…note the large differences in computed linear warming trends (click on plots for high res versions):
The next plot shows the differences between my ISH PDAT dataset and the other 2 datasets. I would be interested to hear opinions from others who have analyzed these data which of the adjustments NOAA performs could have caused the large relative warming in the USHCN data during 1995-97:
From reading the USHCN Version 2 description here, it appears there are really only 2 adjustments made in the USHCN Version 2 data which can substantially impact temperature trends: 1) time of observation (TOB) adjustments, and 2) station change point adjustments based upon rather elaborate statistical intercomparisons between neighboring stations. The 2nd of these is supposed to identify and adjust for changes in instrumentation type, instrument relocation, and UHI effects in the data.
We also see in the above plot that the adjustments made in the CRUTem3 and USHCN datasets are quite different after about 1996, although they converge to about the same answer toward the end of the record.
UHI Effects in the USHCN Station Trends
Just as I did for the ISH PDAT data, I correlated USHCN station temperature trends with station location population density. For all ~1,200 stations together, we see little evidence of residual UHI effects:
The results change somewhat, though, when the U.S. is divided into 6 subregions:
Of the 6 subregions, the 2 with the strongest residual effects are 1) the North-Central U.S., with a tendency for higher population stations to warm the most, and 2) the Southwest U.S., with a rather strong cooling effect with increasing population density. As I have previously noted, this could be the effect of people planting vegetation in a region which is naturally arid. One would think this effect would have been picked up by the USHCN homogenization procedure, but apparently not.
Trend Agreement Between Station Pairs
This is where I got quite a surprise. Since the USHCN data have gone through homogeneity adjustments with comparisons to neighboring stations, I fully expected the USHCN trends from neighboring stations to agree better than station trends from my population-adjusted ISH data.
I compared all station pairs within 200 km of each other to get an estimate of their level of agreement in temperature trends. The following 2 plots show the geographic distribution of the ~280 stations in my ISH dataset, and the ~1200 stations in the USHCN dataset:
I took all station pairs within 200 km of each other in each of these datasets, and computed the average absolute difference in temperature trends for the 1973-2012 period across all pairs. The average station separation in the USHCN and ISH PDAT datasets were nearly identical: 133.2 km for the ISH dataset (643 pairs), and 132.4 km for the USHCN dataset (12,453 pairs).
But the ISH trend pairs had about 15% better agreement (avg. absolute trend difference of 0.143 C/decade) than did the USHCN trend pairs (avg. absolute trend difference of 0.167 C/decade).
Given the amount of work NOAA has put into the USHCN dataset to increase the agreement between neighboring stations, I don’t have an explanation for this result. I have to wonder whether their adjustment procedures added more spurious effects than they removed, at least as far as their impact on temperature trends goes.
And I must admit that those adjustments constituting virtually all of the warming signal in the last 40 years is disconcerting. When “global warming” only shows up after the data are adjusted, one can understand why so many people are suspicious of the adjustments.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Victor, you are right. I failed to mention that after computing the CRUTem3 and USHCN anomalies, I offset them vertically so all 3 datasets averages matched each other over the 1st 3 years (1973-75).
Andrew, the anomalies are relative to 1973 through 2011….but also see my comment above.
Well done, Dr. Roy. A couple of comments. First, you say:
It is not generally realized that good correlation between the data does not mean agreement between the trends. For example, consider the following graphs:
Now, the correlation of all of these is greater than 0.9 … but their trends are all over the map. I discuss this chart further in my post “GISScapades“.
Indeed. My point of view about “adjusted” data is that if you adjust data, your confidence interval must include the original data, the adjusted data, and in addition it must encompass the original data with each adjustment added in separately. In the case where the adjustments are about equal to the final trend, of course, this means that the trend will likely be within the error bars …
w.
What, pray tell, is a legitimate “Station Location Quality” Adjustment, that is (a) not a UHI effect, and (b) is a net positive 0.20 C over 60 years? By this, it means that an average station’s reading today must be RAISED by 0.20 to make it functionally equivalent to the same station 60 years before.
As I understand the history, Anthony Watts, started his sleigh ride by investigating a change from white-wash to white latex paint that would require a 1.00 negative adjustment, not a positive one. Encroaching parking lots? That’s another negative adjustment.
Oh, I know! There are fewer incinerators today than 60 years ago. /sarc
Let’s see that list of all positive and negative site location adjustments that are possible. There number and gross sizes should amount to some staggering statistical error bars.
I am not a fan of any “adjustments” to experimental data.
As far as I am concerned “adjusting data” is the same as making up your results.
Instead what should be done, if climatology is a proper science and if they can not better control their experimental process, is to just increase the stated error for the data set as all of these data quality problems (changing measurement technique, location problems causing UHI, measurement time of day inconsistencies) are effectively instrumentation and measurement errors and should be stated as such. Anything else is just manufacturing results and hiding the real confidence level in the data set.
I did a TempLS run using monthly GHCN unadjusted data, ConUS. This data is as it says, unadjusted – as reported by the met stations. I got a trend 1973-2011 (actually Jan 2012) of 0.161 C/decade. A bit less than CRUTem 3, but not nothing.
In other words, instead of admitting up front that they really do not have useful data on which to draw the kind of conclusions that they are making, due to the poor and inconsistent experimental method used to gather this data, climatology is just making things up and lying. My only conclusion is that field of climatology is currently not a science any more than alchemy is a science.
Nick Stokes says:
April 13, 2012 at 5:48 pm
Nick, how did you average data to avoid overweighting the east coast where there are lots of stations?
Thanks,
w.
Nick Stokes says: April 13, 2012 at 5:48 pm I did a TempLS run using monthly GHCN unadjusted data, ConUS. This data is as it says, unadjusted – as reported by the met stations. I got a trend 1973-2011 (actually Jan 2012) of 0.161 C/decade. A bit less than CRUTem 3, but not nothing.
Nick, many of us have done similar calculations, but the open question is still: Did the Met Station Country Authority adjust the data before sending it to CHGN?
It is actually quite hard to find useful data sets from this country that can confidently be authenticated as “RAW”. If you have a cache, do let us know. Also, can you tell us if this RAW data is the same as received by GHCN?
Nick Stokes says:
April 13, 2012 at 5:48 pm
I did a TempLS run using monthly GHCN unadjusted data, ConUS. This data is as it says, unadjusted – as reported by the met stations. I got a trend 1973-2011 (actually Jan 2012) of 0.161 C/decade. A bit less than CRUTem 3, but not nothing.
You found a trend in a compound metric.
Now do the same but use humidity to calculate the atmospheric enthalpy and then from that the average kilo Joules per kilogram of atmosphere. You may be surprised at what you find as (contrary to the AGW hypothesis) global humidity has been dropping . You will also then be using the correct metric for measuring heat content in a gas. Atmospheric temperature alone is meaningless average global temperature is like an average telephone number.
RE
Willis Eschenbach says:
@ur momisugly April 13, 2012 at 5:21 pm
“My point of view about “adjusted” data is that if you adjust data, your confidence interval must include the original data, the adjusted data, and in addition it must encompass the original data with each adjustment added in separately”.
—————–
Agreed, but of course, you’re preaching to the converted. The official adjusters claim that the unadjusted data are an inherently biased version of reality, and all they’re seeking to do is to remove those biases to provide an uncorrupted version of reality (a Utopian version of reality, some might say).
So then, no need for them to calculate error bars using unadjusted (nasty, biased) data… Isn’t that really what they’re saying? In other words, it is they who decide which reality we get to call reality. It’s purely Orwellian.
On a related question, can I ask if you or Roy to briefly address the idea that linear interpolation of data in the land surface temperature record over time provides an efficient mechanism to propagate and transmit localised spatial-temporal biases (eg. from UHI’s) systematically – throughout the temperature record. Under the cloak of “homogenization”.
And, whatever the expressed justification for it might be, linear interpolation over time will link most if not all data through space and time. The statistical pre-requisite of independence of cannot be not satisfied – invalidating attempts that seek to measure or compare temperature trends through space and time using these data.
The data are not fit for the purposes to which they are directed.
Where haveI gone astray in my thinking? Where are the holes in the argument?
Thanks.
Display note: Pleazze, do not, evah, plot lines or dots or labels in yellow. Really. It’s display screen invisibility ink.
Correction to my point at 6.29pm above.
Should read: …statistical pre-requisite of independence cannot be satisfied…
Sorry.
[Nick Stokes says:
April 13, 2012 at 5:48 pm
I did a TempLS run using monthly GHCN unadjusted data, ConUS. This data is as it says, unadjusted – as reported by the met stations. I got a trend 1973-2011 (actually Jan 2012) of 0.161 C/decade. A bit less than CRUTem 3, but not nothing.]
Do you consider an increased temperature of 1.61C/century by 2073 as CAGW?
Dr. Spencer, thanks for, once again, casting light on this subject. It needs hammered on, over and over again.
I do agree with Brian H, please avoid yellow on graphs if at all possible.
In case you have forgetten, I’d like to remind you that we’re doomed.
RE
Andrew says:
@ur momisugly April 13, 2012 at 6:29 pm
…or indeed anyone who can enlighten me…
PS. I wonder if the land surface temperature records are not more appropriately addressed using statistics better suited to the analysis of neural networks – or other techniques that can accommodate sampling dependencies/ data linkages…?
Gail Combs says: “Finding errors in a computer program is like finding mushrooms in the forest having found one look for others in the same place” ~ A. P Ershov
In my country we say it shorter: “Where bug is, bugs are.”
Although I hadn’t intended the pun, come to think of it, this might have inadvertently hit the mark: the work of fiction known as the land surface temperature record is perhaps better suited to statistical techniques capable of probing the workings of the human brain…
I’m sorry, I know this will seem trollish, but every time I see the good ‘ol average = [(Tmax+Tmin)/2]; I can’t help but to think who ever thought that up wouldn’t have done very well on the TV show “Are You Smarter Than a 5th Grader”.
How do the satellite temps compare? Are these adjusted too? And what about BEST?
Willis Eschenbach says: April 13, 2012 at 5:57 pm
“Nick, how did you average data to avoid overweighting the east coast where there are lots of stations?”
I used inverse density weighting, measured by 5×5° cells. That does a fairly good job of balancing. But prompted by your query, I ran a triangular mesh weighting, which places each station in a unique cell, and weights by the area. Running a monthly model, as with the first example, the trend came down to 0.42°C/decade. But with an annual model, it went back to 0.161.
I’ll write a blog post with more details.
I had assumed that USHCN’S homogenization technique was mostly to increase rural temperatures until they match urban. Then they just generally decrease the overall temperature by a flat 0.004c per decade (the ipcc’s officially sanctioned estimate of UHI).
Apologies fo being O/T [snip . . please repost on Tips & Notes . . kbmod]
Geoff Sherrington says: April 13, 2012 at 6:14 pm
“Nick, many of us have done similar calculations, but the open question is still: Did the Met Station Country Authority adjust the data before sending it to CHGN?”
Unlikely. Most adjustment is done years later, following some perceived discrepancy. But the GHCN data is as recorded within a month or so.
You can see this with data from our town. Within minutes readings are on line. Within hours, they are on the monthly record. And at the end of the month they go on the CLIMAT form. From which they are transcribed directly into GHCN adjusted. You can check. They don’t change.
Even before the internet, the GHCN data was distributed to thousands of people by CD. You can’t adjust that.