Nylo observed something I did not notice:
With this new GISTEMP data release, lots of changes have occurred. There is a generalised warming of the past (until about 1940) and a generalised cooling of the last decade. In total, 91 past monthly records have warmed (most from long ago), and 39 past monthly records have cooled (most within the last decade).
2007 is still the second hottest year on record, but after cooling 0.01C it is now tied with 1998 in that position. 2002 and 2003 have also officially cooled 0.01C.
The overall 1880-2008 warming trend has cooled from 0.56234C/100y to 0.56072C/100y. The 2001-2008 trend has also cooled from -0.1141 to -0.1409. If 2009 stays at its currently high 0.5285C average temperature anomaly, the 2001-2009 trend will still be negative by the end of the year.
Since I had plotted temperatures since 1990 and 2001, I thought I’d show the effect of historic revisions on trends from 1990 and 2001 through June 2009. (Temperature anomalies are multiplied by 100, as reported by GISSTemp). Below, you can see GISSTemp revisions caused the trend from 1990-June 2009 to drop slightly. I usually round to either 0.01C/century or 0.1C/century. (The choice depends on context.) Based on the July update data I might have reported trend as 1.80 C/century; based on the August update, as 1.79 C/century. Otherwise, I might have rounded both to 1.8 C/century; that is, the changes are so small as to make no difference to discussion in the post.
I could show a similar figure for the changes for trends since 2001, but… well… boring. The trend changed from -0.1426 C/century based on the July update on July’s version of Jan 2001-June 2009 and -0.1558 C/century based on the August update. If I’d been in a mood to not round too much, I would have reported -0.14C/century based -0.16 C/century.
Will the trend since 2001 go positive?
Nylo suggest if the 2009 temperature remain at the current 0.5285C average temperature anomaly for the year, the trend at year end will still be negative.
For balance, I’d note that we are in an El Nino now, and it’s predicted to deepen. If the monthly temperature freeze at 0.60C, the least squares trend based on GISSTemp since 2001 will flip to positive when September’s value comes in, and we will end the year with a trend of +0.13C/century based on monthly averages.
I don’t think that the value of zero itself is that important. I consider the difference between projections and observations more interesting. Still, the “zero” threshold does loom large in many minds. So, we’ll see.
Stop saying GISS doesn’t ever revise downward!
These changes are small. But for those who like to slam GISS revisions for never resulting in less warming… well, it did this time. At least the revisions made the trends drop over periods I examine. 🙂
Lucia,
As far as the “zero” value goes…an observed trend of zero does not fall outside of the 95% range of the model ensemble projected trends until trends of length 10-11 years (i.e. at 10 years the ensemble 95% range (as defined by 2.5 and 97.5 percentiles) includes zero, at 11 years, it does not). I think this is similar to what Easterling and Wehner reported.
So, if “zero” is interesting for some reason, then starting in January 2001 we have to wait until at least Dec 2010 before it (an observed trend of zero) lies outside of the 95% range of model expectations (under SRES A1B). Perhaps the wait would be shorter under models run with A1FI.
-Chip
It actually is interesting to see a drop in GISS trends, however slight they may be.
Interesting that the trend is roughly 0.56C/100yrs. Easy to speculate that errors in the raw thermometry, subsequent corrections, and method of arriving at a global number could combine to create a slope error in the same ballpark.
In an earlier thread I commented the current warm spurt (based on satellite) doesn’t appear to be related to El Nino, but rather is a result of regional conditions, a “belching” of heat if you will. I still think that and rather than intensifying of heat accumulation, it will oscillate and by year’s end or early 2010, El Nino will fade if it even makes the grade. 2007 saw a similar fate.
IMO, we are witnessing further losses of OHC that appears to come in waves manifesting itself in SST, but the heat loss is not being replaced below the surface. It took many years to accumulate heat in the oceans and will take many years for it to diminish, ever so slowly. I think the negative PDO regime change is just getting started.
RPS made this comment recently:
http://climatesci.org/2009/08/11/july-lower-tropospheric-temperature-anomalies/
“What I wanted to comment on here is the clear absence of an El Niño temperature signal in the spatial map of the July anomalies.”
They do it every month, but I never be cease to be amazed at the changes to previous temperatures, especially decades old temperatures…
What have we learned the the last month that allows us to change the temperatures from 50 or 100 years ago?
David Jay,
I think that once a person crosses the threshold into believing in the religion of Climate Science, ‘after the fact’ changes that The Trend chooses to make are just The Trend Being The Trend. (shoulder shrug)
Like when Manny (in the outfield) dove and cut off that throw to the infield from Johnny Damon (20 ft away, also in the outfield)… Manny bein’ Manny. I will never forget that. 😉
Andrew
Chip–
Yes. If zero is interesting per se, and the test is against the distribution of all model runs, that’s what we get.
(I need to figure out why your comments always end up moderated!)
Lucia,
(I need to figure out why your comments always end up moderated!)
Clearly, because your computer has been well-trained! ;^)
-Chip
I white listed your IP. That seems to have fixed things.
I have just seen the NASA anomaly temp map for July. Note Britain showing up in the 1C – 2C above average band. Yet their own Met Office reports that July was about average with only parts of Scotland being about 0.8C above average.
Something stinks. Any other ‘anomalies’?
David Jay (Comment#17758) August 12th, 2009 at 11:17 am queried: “What have we learned the the last month that allows us to change the temperatures from 50 or 100 years ago?”
My understanding is that when the time series of a station is adjusted for something like a change in instrument type, or change in the time of reading, or some undocumented change detected by a homogeneity check that the normal way to apply the correction is to apply it to the historical data.
The philosophy is to consider the current reading of the instruments to be the “true” reading and then to adjust the older readings in such a way as to have the most accurate trends.
So if somehow there is a sudden unexplained step change upward of 0.5C compared to all surrounding stations, then it would be assumed that something has changed at that station. Either its exposure or in the instruments.
So now that you know that the station data has made an unexplained 0.5C jump that is inconsistent with surrounding stations the question is how to correct for it.
One method would be to continue recording that station’s info as read on the thermometer, but to go back and move the past (adjusted) data upward by 0.5C. This prevents the unexplained 0.5C jump from having any effect on temperature trends calculated from that station’s historical record. It does change the absolute temperatures in the historical record.
What this means is that the complete set of data for a station will have at least two time series. One set is the initial raw readings actually reported. Another set will be ones that have been adjusted.
See comment 36 on this post http://www.climateaudit.org/?p=6767 to see a chart of 6 different temperature histories for on station, Mt. Gambier, South Australia.
GISS antarctic warmer?
000
NOCA42 TJSJ 011628
PNSSJU
PUBLIC INFORMATION STATEMENT
NATIONAL WEATHER SERVICE SAN JUAN PR
1228 PM AST SAT AUG 1 2009
…SAYING GOOD BYE TO A VERY UNUSUAL MONTH…JULY 2009…
WHILE THE EASTERN THIRD OF THE UNITED STATES WAS BUSY POSTING 1,788
RECORD COLD TEMPERATURES…SAN JUAN WAS UNUSUALLY WARM. IN
FACT…JULY 2009 WAS THE THIRD WARMEST JULY EVER IN SAN JUAN…WITH
AN AVERAGE OF 84.8 DEGREES. THIS WAS JUST BEHIND THE WARMEST JULY
IN 1980 WITH 85.1 DEGREES AND THE SECOND WARMEST JULY WITH 84.9
DEGREES THE YEAR AFTER IN 1981. BUT IT WAS A MONTH IN WHICH THE
AVERAGE MINIMUM TEMPERATURE WAS THE WARMEST EVER RECORDED…WITH
79.5 DEGREES…BREAKING THE PREVIOUS RECORD OF 79.0 DEGREES SET IN
1981. ALTHOUGH THE MAXIMUM TEMPERATURES IN PUERTO RICO WERE BY NO
MEANS THE HIGHEST…SAN JUAN NEVER RECORDED MORE THAN 93 DEGREES
LAST MONTH AND THAT HAPPENED ONLY ONCE ON THE FIRST…IT WAS THE
STEADY STREAM OF DAYS WITH 90 DEGREES AND ABOVE…22 IN ALL…THAT
CONTRIBUTED TO MAKING EVEN THE AVERAGE MAXIMUM TEMPERATURE FOR
JULY TO BE THE SIXTH HIGHEST SINCE RECORD KEEPING BEGAN IN
1850…WITH EXACTLY 90 DEGREES. EVEN THE DAY WITH A RECORD
RAINFALL OF 2.31 INCHES…THE TWENTY FIRST…HAD A HIGH OF 90
DEGREES AND NO DAY WAS LESS THAN 88 DEGREES.
PERHAPS WE CAN BLAME IT ON EL NINO WHICH BEGAN TO CHANGE GLOBAL
WIND FLOW PATTERNS OR ON THE JET STREAM WHICH DIPPED TO VERY LOW-
LATITUDES SO LATE IN THE SUMMER OVER THE EASTERN UNITED
STATES WITH HIGH PRESSURE OVER THE WEST COAST BUT…WAS IT GLOBAL
WARMING?
IF YOU WERE IN SAN ANTONIO TEXAS YOU MIGHT BELIEVE IT…THEY HAD
THEIR HOTTEST MONTH EVER…NOT JUST JULY…WITH 22 DAYS REACHING
OR EXCEEDING 100 DEGREES LAST MONTH. YESTERDAY PORTLAND OREGON
WAS SET TO HAVE ITS SECOND HOTTEST JULY EVER…AND SEATTLE HAD
THEIR HOTTEST DAY EVER LAST MONTH.
ON THE OTHER HAND…YOU MIGHT MEET SOMEONE FROM FROM CHICAGO
WHERE THE AVERAGE TEMPERATURE WAS 68.9 MAKING IT THE COOLEST JULY
SINCE 1942 WHEN THE STATION WAS MOVED AWAY FROM THE LAKEFRONT…OR
GRAND RAPIDS MICHIGAN THAT BROKE THEIR RECORD COOLEST JULY SET IN
1992. OR MORE LIKELY SOMEONE FROM NEW YORK CITY WHICH HAD ITS
SECOND COLDEST JULY. YOU MIGHT EVEN MEET SOMEONE FROM THE SOUTH
POLE WHICH HAD A BONE CHILLING AVERAGE TEMPERATURE OF MINUS 86.8
DEGREES FAHRENHEIT BREAKING THE RECORD SET BACK IN JULY 1965.
THEIR LOWEST TEMPERATURE LAST MONTH WAS ONLY 108.2 DEGREES…BELOW
ZERO. IT MAKES OUR MID 80S AVERAGE LOOK POSITIVELY INVITING.
$$
SNELL
Carlo–
Someone ought to tell the National Weather Service to buy a keyboard with lower case letters. 🙂
Hi Lucia
We have confirmed a warm bias in the multi-decadal global average surface temperature trend (which in the only estimate we have can explain about 30% of the IPCC reported trend); see
Klotzbach, P.J., R.A. Pielke Sr., R.A. Pielke Jr., J.R. Christy, and R.T. McNider, 2009: An alternative explanation for differential temperature trends at the surface and in the lower troposphere. J. Geophys. Res., in press.
http://www.climatesci.org/publications/pdf/R-345.pdf
Best Regards
Roger Sr.
The weather service bulletin above was all caps because it was prepared and sent out via an ASR 33 teletype, which only has 32 upper case characters, plus another 32 characters/numbers/symbols that can be activated by sending a shift character.
I’m kidding of course, but the original baudot code 10 character per second mechanical teletypes is what set the standard data formats still used in a lot of cases, such as aeronautical weather METARS reports.
Chip Knappenberger (Comment#17752): Ah, but there has been no warming in UAH’s LT since April 1997! I bet THAT is outside someone’s confidence intervals…