An odd twist has developed in the past week regarding some data sets that surfacestations.org volunteers have been using to look at individual stations. The data has changed on NASA’s GISS website with no notice whatsoever.
My first indication that something changed came from surfacestations.org volunteer Chris Dunn who wrote to me complaining that one of the sites he’d recently surveyed, Walhalla, SC had been greatly adjusted at GISS for no good reason that he could ascertain, since the site is pristine by climate monitoring standards, and has not gone through any significant changes in the past, and has been operated at the same location (by the same family) since 1916. He wondered why NASA would have to adjust the data for a “good” station. The way I view it, shouldn’t good data stand on it’s own? That was September 7th. He was using data from NASA GISS published on 8/28.
So he continued to look at the data, and the site. The on Sept 11th he noticed a change when he downloaded the data again. Something had changed, the data was different. Not only the adjusted data but the “raw” data too.
Steve McIntyre of Climate Audit has a complete review at: http://www.climateaudit.org/?p=2077 where he traces data back to Detroit Lakes, MN the station that started this all. See my original post on this: http://www.norcalblogs.com/watts/2007/08/1998_no_longer_the_hottest_yea.html
This set other people into motion looking at the NASA GISS data sets. The conclusion? NASA published new raw and adjusted data on their website with no formal or informal notice. I don’t know what to make of this, by I think perhaps this could be a breach of the Data Quality Act. At the least, it flies in the face of accepted scientific courtesy, where if you publish data sets being used by researchers worldwide, scientific courtesy would dictate that you at least place notice of such a change, otherwise there can be a domino effect for hundreds of research projects that use the data. Which would cause researchers to wonder why things don’t look the same anymore and begin searching for answers. Well that is exactly what happened here. We had a citizen trying to figure out why a climate site with good data was “adjusted”, and then the data changed right in the middle of him looking at it.
Whether this was accidental or intentional I cannot say, but it certainly does not look good coming on the heels of NASA GISS’s most recent issue of a mistake causing a revision of our temperature history on August 8th. We deserve better accounting than this when so much hinges on this data.
Let’s give NASA and Hansen the benefit of the doubt and see what they have to say about it.
UPDATE: NASA has posted today, their explanation which you can read here: http://data.giss.nasa.gov/gistemp/ Note that this notice appears a full week after the data changed (about 9/10) and only after there was discussion of the issue on blogs such as Climate Audit over the weekend. Why would NASA GISS not announce the change at the same time the data did, particularly when the announcment of the change ammounted to one small paragraph?
There are indications that there are now more than 1221 stations used by GISS. Should they be added to the survey?
note that the database has two Allegan, MI bins.
***NOTE FROM ANTHONY: those other stations in GISS have always been there and I’m well aware of them…had to start someplace, so chose USHCN set first.
I do not get it. These people (Nasa) claim to be engineers and scientist but can not do the most fundamental data control? As a software engineer we use version control systems to allow us to track changes and notify others. Thanks for the great work guys.
I especially like their comment:
The effect of station removal on analyzed global temperature is very small, as shown by graphs and maps available here.
They’ve already said that the corrections to the U.S. station data had no real effect on ROW readings (the U.S. being 2% of land mass). Yet they still correct the data.
Now removal of U.S. stations has no effect. How many earth stations would have to be removed before there IS an effect?
Does anyone else get the impression that there is about to be a great unraveling of the agenda-driven “scientific consensus”?
For a moment I thought I had taken crazy pills.
When Anthony first posted Orland and Marysville I got interested. It was simple, I’ve travelled through the areas for the past 10 years.It struck a
chord.
The very first thing I did was this.
Download Orland TMAX and TMIN. Download Marysville TMAX and TMIN:DAILY data
from USHCN. Why daily?
Its turtles all the way down. Start at the bottom.
Then I compared. Not which was “hotter”, that’s dumb. but did one site show a bigger increase than the other? Orland was getting warmer, BUT Marysville was getting warmer than warmer. So, I checked other sites around Orland, other sites around Marysville and it seemed pretty darn clear. Marysville wasnt a good or representative site.
Now… that needs explaining. ( if you’ve driven down 99 in the YUBA city/Marysville area it NEEDS No explaination.. )
1. Did Orland warm? Yes.
2. Was Marysville worse? Yes.
NOW, when I go back to replicate those early studies the data is SERIOUSLY FUBAR AND DIFFERENT!
History is being revised.
You thought it was 13C on
August 1, 1967. Well, Today, we changed our minds. It was 13.2C. errrr
wait.. it was 13.6C… errr wait it was 12.9C.
WHAT YOU SEE IS HISTORICAL RELATIVISM BEFORE YOUR VERY EYES.
Anyways, I went back to check my work, from brick # one
Same sites. same url…. Some data files have been seriously corrupted. If it were a .1C here or .2C there I would take a different tone. But the daily data has been seriously eff up.
Its like catching Sandy Berger mid theft.
“Let’s give NASA and Hansen the benefit of the doubt”: let’s not. Let’s recognise that the default interpretatation is that there’s dirty work afoot.