Guest Essay By Walter Dnes
There is much interest in the latest temperature anomaly adjustments by NOAA/NCEI (formerly known as NOAA/NCDC). This author has been downloading NOAA monthly temperature anomaly data since 2010. The May 2015 adjustment is not the only one. There appear to have been 8 adjustments between November 2010 and May 2015. Assuming that these changes are legitimate adjustments, one has to wonder, if they got it wrong the last 7 tries, what confidence can we have that they got it right THIS TIME, or will they change it again if Earth doesn’t cooperate? To paraphrase Darth Vader…
Credit and a special thanks to Josh for his incredible artistry and humor!
The NOAA/NCEI monthly raw datasets from January 2010 to May 2015 have been uploaded on WUWT to here for those of you who might wish to do your own analysis. I’ve also included some data documentation in the readme.txt file included in the download. Current NOAA/NCEI data can be downloaded here, click on “Anomalies and Index Data”.
There are 65 months from January 2010 to May 2015. Eight of those months saw significant changes in the anomaly data. There were only very minor changes from January 2010 to October 2010. Note also that the data was originally available to both 2 and 4 significant digits. It is now available to only 2 significant digits. The 2-digit values appear to be rounded-off versions of the 4-digit data.
- The first notable change occurred in November 2010, with most anomalies adjusted upwards over the period of record. Mid 1939 to mid 1946 was not raised. Keeping it unchanged while everything else is bumped up is effectively equivalent to lowering it. Of interest is that for the period 1880-to-1909, anomalies for the two months April and November received the most significant boosts.
Walter Dnes – Click the pic to view at source - The next change occurs in April 2011. The period 1912-to-1946 appears to be depressed relative to the rest of the record. Here is the delta between March 2011 and April 2011.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to April 2011.
Walter Dnes – Click the pic to view at source - The next change occurs in October 2011. The periods 1880-to-1885 and 1918-to-1950 appear to be depressed relative to the rest of the record. Here is the delta between September 2011 and October 2011.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to October 2011.
Walter Dnes – Click the pic to view at source - The next change occurs in January 2012. The period 1905-to-1943 appears to be depressed relative to the rest of the record. 1974-onwards is raised relative to the rest of the record. Here is the delta between December 2011 and January 2012.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to January 2012.
Walter Dnes – Click the pic to view at source - The next change occurs in February 2012. The period 1898-to-1930 is raised relative to the rest of the record. Here is the delta between January 2012 and February 2012.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to February 2012.
Walter Dnes – Click the pic to view at source - The next change occurs in August 2012. The period 1880-to-1947 is lower relative to 1948-to-2010. Here is the delta between July 2012 and August 2012.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to August 2012.
Walter Dnes – Click the pic to view at source - The next graph is not an adjustment. It’s a sanity check. February 2014 was the last available month of 4-significant-digit data. Starting March 2014, 2-significant-digit data is being used. The comparison between February 2014 and March 2014 confirms that the 2-digit data is a rounded-off version of the 4-digit data. The “jitter” is within +/- 0.01, i.e. roundoff error.
Walter Dnes – Click the pic to view at source - The next change occurs in April 2015. The 2-digit data results in a more jagged, sawtooth graph. The period 1880-to-1905 is slightly raised, and the period 1931-to-1958 is slightly lowered relative to the rest of the record. Here is the delta between March 2015 and April 2015.
Walter Dnes – Click the pic to view at source And here is the accumulated change from October 2010 to April 2015.
Walter Dnes – Click the pic to view at source - Some of the changes from April 2015 data (downloaded mid-May) to May 2015 (downloaded mid-June) look rather wild. A drop of as much as 0.14 C degree in 1939 anomalies and a rise of as much as 0.15 C degree in 1945 anomalies made me do a double-take, and inspect the data manually to insure there was no error in my graph. The raw data confirms what the spreadsheet graph shows…
Data for April 2015 versus May 2015 Data month April 2015 May 2015 Change 1938/11 0.11 0.01 -0.10 1948/12 -0.13 -0.25 -0.12 1939/01 -0.02 -0.16 -0.14 1939/02 0.01 -0.11 -0.12 1944/12 -0.02 0.10 0.12 1945/01 0.01 0.16 0.15 1945/02 -0.13 0.02 0.15
And now for the “pause-buster” adjustment. Here is the delta between April 2015 and May 2015. This adjustment is a roller-coaster ride.
- The period 1880-to-1925 is up-and-down
- 1926-to-1937 is relatively stable, down approximately 0.03 to 0.04 degree from April.
- 1938-to-1939 crashes down to 0.10 degree below April.
- The adjustment spikes sharply up to +0.15 by the end of 1944
- It drops down sharply to 1948.
- Slides gradually down to 1963.
- Stable 1963-to-1973
- Rises 1973-to-1980
- Stable 1980-to-1992
- Falls 1992-to-1998
- Rises 1999 to November 2010 (end of comparison)
And here is the accumulated change from October 2010 to May 2015. Because the May 2015 change is the largest, the accumulated change from October 2010 to May 2015 is similar to the May 2015 monthly change. One thing that stands out… because 5 or 6 of the 8 adjustments pushed down part or all of the years between WWI and WWII, there is a marked drop from 1920 to 1939 in adjusted temperatures. This has the effect of doing to “The Dirty Thirties” what Michael Mann tried to do the Medieval Optimum warm period; i.e. erasing it from the temperature records. So as our friend Daft Bladder says, “I am altering the data. Pray I don’t alter it any further”.
It’s all about ‘Oversights’
“May 15, 2015: Due to an oversight several Antarctic stations were excluded from the analysis on May 13, 2015. The analysis was repeated today after including those stations”.
http://data.giss.nasa.gov/gistemp/updates_v3/
Sarc.
“I still cannot see how the UHI affects temperature measurements over the oceans and in remote areas.”
Read up on station temperature spatial gridding. Stations that do exist (in hotter UHI areas) “spread” their temperatures into areas where the temperature is unknown and thus assumed. They killed off many of the stations that were in cooler areas.
Yes and keep in mind the vast majority of the human population live near the coasts… making UHI that much worse over the oceans then land areas.
climanrecon (July 9, 2015 at 5:59 am) suggests that
It does however seem unreasonable that these past numbers should change quite drastically overnight, then in a short time again change quite drastically overnight, then in another short time …
These images are part of a blog post still under construction, examining the behaviour of GHCN-M adjustments for past temperatures at Marseille and nearby (in the climate data sense) stations, using saved ghcnm data sets. The choice of Marseille arises from a blog post LE GISS ET LES SÉRIES LONGUES DE TEMPÉRATURES. I have seen similar behaviour closer to home with past data for Irish stations, but illustration using an Irish station would have restricted the choice of nearby stations to a generally easterly direction, while choice of Marseille provides nearby stations around the compass in France, Spain, Italy and Switzerland. There is no special reason for choosing the past temperature values for January 1978 – I simply took a year from the plotted temperature records for Marseille in that blog post, and used that year for other nearby stations as well. I have pointed out to the owners of that blog (the post itself did not provide an opportunity to add comments) that GISS are using the adjusted GHCN-M data as input, and that this, rather than the Gistemp processing, may be the source of the variations discussed in the blog post.
For anyone interested I will publish the blog post in the next day or two once I have added a little more detail, in its still “under-construction” state, then complete the post later. The link will be https://oneillp.wordpress.com/2015/06/23/marseille-jan-1978/, but this link will not be accessible until I have published it, sometime today or Saturday. I will add another comment here when it becomes accessible.
For harrytwinotter (July 9, 2015 at 5:51 am):
The code on the NOAA website appears to be v3.0.0, not the code currently used. I was tempted to download and run this code to try to determine the cause of these erratic adjustments, but thought better of it in the absence of current code. Having downloaded and recoded the Gistemp code with additional diagnostic output, I am aware of the scale of such an undertaking. It may come as a surprise to Harry to find that some of us have “had the energy” to do this, and have contributed by notifying GISS of bugs found in their code – another good reason for making the code available. You can verify that I have done so by looking for my name at http://data.giss.nasa.gov/gistemp/updates/
My images seem to have been lost. the URLs are http://wp.me/aQRhu-Gy and http://wp.me/aQRhu-GV
I think I see why they were lost. Trying again:
in GHCN-M v3″ width=”” />
Note how the January 1978 changes frequently in GHCN-M v3, even overnight, and that the changes include correction for both “urban heating” and “urban cooling”. Anyone willing to suggest station relocations take place this frequently? Problems with Pairwise Homogeneity Adjustment seem more likely.
The detailed blog post is coming soon
And again:
[blank? .mod]
Based on your latest comment, I’m gonna predict that you can’t be wrong.
=============
My blog post is now published at Wanderings of a Marseille January 1978 temperature, according to GHCN-M
I’ll have one more try here posting an image to show the relevance of the post to this topic.
If this fails again,and a moderator sees this (all this comment including image shows correctly in preview in Waterfox before posting – but I see no comment toolbar or preview in Chrome or IE?): the img code above is (left-delimiter)img src=”https://oneillp.files.wordpress.com/2015/06/marseillejan783.jpeg” alt=”Marseille January 1978 temperature” width=”” /(right delimiter), and I’d be grateful if you could reinsert it for me.
The post link is (left-delimiter)a href=”http://wp.me/pQRhu-Gm”(right-delimiter)Wanderings of a Marseille January 1978 temperature, according to GHCN-M(left-delimiter)/a(right-delimiter) if that fails too!
And for some WordPress reason I fail to understand, clicking on the images in those two links to see the full-size images brings up images for other stations in the blog post I am working on instead of the full-size images. Click on the underlined size (1920 x 978) to get the full-size images.
These should have appeared between “then in another short time …” and “These images are part of a blog post still under construction” near the start of my first comment.
To be blunt: It appears the most logical treatment for all these adjustments is to consider them finagle factors promulgated by disciples of Professor Lyon E, Zossov.
Wow! Almost like data hacking.
Did NOAA predict this
They may not need to alter it further. It looks like they do not have much more to prove. You may recall a former quote from NOAA:
”The simulations rule out (at the 95% level) zero trends for intervals of 15 yr or more, suggesting that an observed absence of warming of this duration is needed to create a discrepancy with the expected present-day warming rate.”
Guess how the latest changes affect things? As a result of exchanges in my earlier post on NOAA, I was prompted to check out things on Nick’s site and this is what I found.
Temperature Anomaly trend
Jun 2000 to May 2015
Rate: 1.246°C/Century;
CI from 0.597 to 1.894;
And the first time the lower end of Cl is negative is:
Temperature Anomaly trend
Feb 2009 to May 2015
Rate: 2.372°C/Century;
CI from -0.030 to 4.774;
(Keep in mind that all other times for the other data sets was no statistically significant warming for between 14 and 22 years.)