Is warming on 11 year hiatus? Well… I always think it’s best to ask yourself: Do I really think a particular method of looking at data is meaningful? Would I still believe this if the answers turned out “wrong” from my POV? Or, will I eventually find myself explaining my own method gives uncertainty bounds that are “too small”, when my method suddenly gives “wrong” (in my eyes) answers?
So, in this regard, I need to ask Tilo: Why throw out GISS Temp?
I know there are people who love to hate GISS Temp. I understand the arguments both for and against it. Still, I think it’s probably a bit premature to be decreeing warming is on an 11 year hiatus by ignoring GISS Temp and NOAA/NCDC. Here’s what I get for the 12×11 = 132 month trends:
Click for larger.Compare and contrast to Tilo Reber’s graph:
As I see it: When the graph of the average of all five measurement groups goes flat, then one can claim an 11 year flat trend. One will be able to claim it as long as that situation sticks.
But throwing out GISS and NOAA? That’s not quite fair!
I’d point out that NOAA USHCN data is the basis data for both HadCRUT and GISS, nobody is ‘throwing it out” it s there, as the base data, just like the base data from NOAA in the satellite record is used for UAH and RSS.
It gets considered, just not top billing because it is used to create the other four end products.
Ok– Fair enough on NOAA. I include it because it does give different results than the other five.
Right now GISS is the one really driving positive trend for the average.
That said, unless the temperature anomalies rise, UAH will probably tick negative on the 132 month next month. In fact, even though GISS is the driver for the poistive trend, I wouldn’t be surprised at all if we go flat by my measure next month. But, the coolth has to hold a bit longer to get all five to go flat or negative.
Lucia, in my business “bad data” is usually found and thrown out immediately. This is especially true when you can sift through the data carefully as Watts up and CA have done. Garbage in garbage out. Including bad data in the mix only confuses the decision making process. This is especially true when there is indeed confusion stemming from the bad data. Warming activists love to quote GISS temp for example of having March 2008 as a very warm month of 2007 as the warmest year. Both are not true but allow policy making decisions to be made incorrectly.
Bob B- “not true” may be going a bit far. If a claim is dubious and contradicted, but not disproven, by other evidence, I tend to say “debatable” and strongly hint which conclusion is probably more accurate. 😉
I’m not sure I agree with lucia that the average is best, as TLT and GMST are not meant to mix. one could easily argue that you should include radiosondes to, and on and on. In fact, don’t some radiosonde data sets have surface values? I think I’ll put together a more comprehensive look at the different measurements. Say, what pressures does LT represent, anyway? HadAT only reports series at different pressure levels.
http://hadobs.metoffice.com/hadat/hadat2/hadat2_monthly_global_mean.txt
Ah! MSU equivalents! I won’t be needing those pressure levels after all. For completeness, here’s the HadAT MSU equivalents:
http://hadobs.metoffice.com/hadat/msu/anomalies/hadat_msu_global_mean.txt
There are other datasets, but these are difficult/impossible to compare with others becuase RATPAC is broken into, at most, seasonal timeseries, and RAOBCORE has a difficult website to use, as does INGRA.
Well… I can compare those later this month. Tomorrow, I’ll be showing the current trends based on the merge and individual data sets.
I agree the satellite data isn’t technically measuring the same thing as the rest. But, it’s been pretty close since the beginning of measurements. I do show everything all together too– and if someone’s point of view really hinged on “pure surface” only, I’d show the NOAA/HadCrut/GISS merge together.
In the long run, it’s difficult for one of the sets to veer off too much without things looking really weird.
Bob B–
I think Steve and Anthony show enough to suggest that one might have reason to suspect the GISS data, or to believe the uncertainty is quite high. But I don’t think they’ve shown the GISS data is obviously worse than the other data. I agree that bad data may be worse than no data. But if the data are bad, I’m not convinced throwing out GISS solves anything.
Is HadCrut obviously better? The satellites had issues when they were new. Are they better than the land measurements? If yes, do we know this for sure? I think we don’t know.
I’ve been working with the HadAT data, and I may be doing this wrong, but the results I get are at odds with the MSU and the surface, showing alot of warming (relatively). Suspicious…
Also, HadAT ends in March and is very noisy.
Andrew–What’s HadAt based on?
Radiosondes:
http://hadobs.metoffice.com/hadat/
Lucia, I think Anthony and Steve have demonstrated the issues UHI and others with the GISS temp set and methodologies. I think if it does indeed continue cooling it would be the proverbial “icing on the cake” for you to show that GISS shows cooling as well. So please do continue using it. I am just very concerned because what is at stake here and bad policy decision made off the constant GISS press releases and people quoting it as gospel.
BTW continue the great job!
Bob B–
It’s not my goal to prove cooling. But your observation about “icing in the cake” is important. You see even GISS showing something as “icing” with systems like UAH as “the cake”. There are others who see GISS as the cake and UAH as “the icing”.
Which ever is which, I’ll always show both the cake and the icing! 🙂
lucia,
Are you familiar with loess fit regression? If you try it with GISS, you’ll find it in the global warming on hiatus camp. I’ve crafted a quick plot you can look at here:
http://i25.tinypic.com/hspbvs.jpg
I use the same 11 year period under discussion, with the GISS anomaly normalized to the 11 year mean. At the end of the period, the loess fit is ever so slightly less than at the beginning, indicating just a hint of downward trend, but not enough that we couldn’t ignore it if we chose to, and called the trend “flat.”
For HadCRUT, UAH and RSS, for the same period, the loess fits all show a bit more marked declines, in starting higher, and ending lower. Here’s a spaghetti graph of the four of them:
http://s3.tinypic.com/2607ibs.jpg
Will you now reconsider that global warming is on hiatus, at least for the past 11 years, even using GISS?
Basil
Basil–
Oh the temperatures have flattened. It’s just not flat for 11 years using OLS for GISS.
We’ll see if it stays flat. I suspect it will start to rise but…. my expectations aren’t data.
Climate audit has been engaging in lengthy analysis of how the GISS is calculated. If you delve through the process, it is not such as to give one any reason to believe the numbers have any meaning.
“So, in this regard, I need to ask Tilo: Why throw out GISS Temp?”
Lucia, I only just saw the issue, and I responded on Anthony’s site. If you don’t mind, I will simply lift it and reproduce it here.
“Basil:
While I respect Lucia’s opinion on which data sets to use, in my mind one has to draw a quality standard somewhere. On top of all of the quality problems that Anthony has documented with the data, I think that there have been adjustment choices made with GISS that are designed to give warming results that increase the trend. I think that this paper does a good job of explaining some of those choices.
http://www.friendsofscience.org/assets/documents/CorrectCorrections.pdf
crosspatch also has a good take on it above.
And of course the divergence over the last decade should also tell you something. The decadal divergence is .13C. That is well over half of the .2C decadal trend that is blamed on CO2 by the warmers.
Once we do a better analysis on HadCru data we may find out that it has many problems of its own. But for right now I have to draw a line in the sand somewhere, and my personal choice is not to accept GISS for any purpose. I believe that it would be excusing agenda oriented work to use it. Just my opinion.”
Tilo–
Sorry, I thought I linked you too, and you’d see a track back!
As I said: I do understand why people would suspect the GISS data. To some extent, I think that even if the GISS algorithm is designed to give a warming trend, this fiddling is self limiting. At a certain point, all the weather services, normal human beings etc. know that while it did warm from the 70 – 2000, it really hasn’t warmed recently.
Even without thermometers, I garden. I know spring was delayed. I have photographs of my gardens, and I know I don’t simply mis-remember when peonies or daffodils bloomed in other years.
On the GISS algorithm, I think my favorite all time comment appeared on climate audit yesterday:
It may well be that the method doesn’t introduce any bias in the mean trend. Given some aspects like using April data to correct march data, it may smooth out weather noise, change spectral characteristics of “(weather!)”. This complicates any advanced statistical analysis, but might leave the mean trend intact.
Anyway, I’m pretty much going to use it along with the rest– even though it does appear demented. Meanwhile, I’m interested in seeing what everyone gets using their choices and judgement. (But, I’ll still comment. 🙂 )
There are a lot of posts here about how long is long enough, so I’m not sure where the question belongs.
My understanding is that The Official Answer is about three decades, more or less.
That being the case, how can the GCM calculations of the effects of Pinatubo, those covered five years, be considered valid calculations? Was the ‘weather noise’ somehow suppressed in the calculations? Were the effects determined by ensemble averaging several runs by several models using several different ICs?
Maybe this has already been covered somewhere nearby.
Thanks