GISS vs CRU/RSS/UAH

[skierinvermont]

1) You misread me. I was talking about the surface/satellite divergence.

2) Because the Arctic results don't answer any questions about GISS extrapolations or the surface/satellite divergence. And because you previously chose to harp on the inherent "innaccuracy and flaws" of UAH, and basically dismissed its relevancy on several occasions. But on other occasions that serve your points, UAH data works just fine for you. It just doesn't make sense, after you inecessantly attacked the people who run UAH (basically accusing them of running from scientific advancement and falsely defending their methods) as well as the methods themselves, to turn around and happily use their data without questioning it.

I am using it as a reasonable approximation for areas where we have no other data. I don't need it to be precisely accurate to do that.

At the same time I do not believe UAH to be precisely accurate globally. I believe its TLT data to be biased low by a few hundredths C/decade. This does not prevent me from using it as a reasonable approximation of arctic temperatures. The magnitude of the arctic trend is nearly two orders of magnitude greater than the potential error in UAH.

There is no contradiction. I believe UAH to be biased low globally (based off of radiosonde data and STAR) but I also believe it gives us a reasonable approximation of arctic temperatures where we have little other data.

[skierinvermont]

FYI here is starting in 2000... only because you insist.

UAH infilling of the poles on either GISS 60-60 or HadCRUT 60-60 causes them to match GISS global this decade. No matter how you slice it. Relative year-year changes and trends are not altered by changes in start point or base period.

[tacoman25]

I am using it as a reasonable approximation for areas where we have no other data. I don't need it to be precisely accurate to do that.

At the same time I do not believe UAH to be precisely accurate globally. I believe its TLT data to be biased low by a few hundredths C/decade. This does not prevent me from using it as a reasonable approximation of arctic temperatures. The magnitude of the arctic trend is nearly w orders of magnitude greater than the potential error in UAH.

There is no contradiction. I believe UAH to be biased low globally (based off of radiosonde data and STAR) but I also believe it gives us a reasonable approximation of arctic temperatures where we have little other data.

I would argue that if it is a reasonable approximation of Arctic temperatures, it is also a reasonable approximation of global temperatures. Magnitude of trends aside, it is still the same methods/data being used.

[skierinvermont]

I would argue that if it is a reasonable approximation of Arctic temperatures, it is also a reasonable approximation of global temperatures. Magnitude of trends aside, it is still the same methods/data being used.

Indeed it is a reasonable approximation of global temperatures. Probably a mere .03 -.04C/decade too cold. More than accurate enough to corroborate GISS's rapid arctic warming.

[tacoman25]

FYI here is starting in 2000... only because you insist.

UAH infilling of the poles on either GISS 60-60 or HadCRUT 60-60 causes them to match GISS global this decade. No matter how you slice it. Relative year-year changes and trends are not altered by changes in start point or base period.

I apologize if you already explained, but why the 1990-99 base? Also, what happened to HadCRU 60/60 with UAH poles? I thought that was on the other graph? I believe if you run trend lines comparing HadCRU 60/60/UAH poles vs. GISS over the past decade, you will see the increasing divergence Will and I have been referring to.

[tacoman25]

Indeed it is a reasonable approximation of global temperatures. Probably a mere .03 -.04C/decade too cold.

Which would then indicate your non-stop lambasting of UAH was a bit overdone. Just sayin.

[skierinvermont]

I apologize if you already explained, but why the 1990-99 base? Also, what happened to HadCRU 60/60 with UAH poles? I thought that was on the other graph? I believe if you run trend lines comparing HadCRU 60/60/UAH poles vs. GISS over the past decade, you will see the increasing divergence Will and I have been referring to.

HadCRUT 60-60 w/ UAH poles is on the other graph.. I didn't include it in that last one because I already had 4 lines and it starts to look cluttered. If you want a specific comparison I can make it pretty quick now that I have all the data in excel.. thinking and making the formulas is what takes time.

I actually did the 1990-1999 base because that allows us to focus on any divergence that has occurred since then (IE over the last decade). I can't make the base a single year because there is a fair amount of noise in single year values. You need at least 10 years I think. And I didn't want to leave it as 1961-1990 because then we would be looking at divergences that have occurred since then. Anyways, it is the year to year changes and trends that are important, which are unaltered by base period selection.

For HadCRUT 60/60 UAH poles vs GISS we do find some divergence if we start in 2000.. you can see that quite clearly if you just look at my 1990-2010 graph of the two and look only at 2000-present (see below). Just look only at the 2000-2010 period and you see there has been some divergence. I actually already pointed out already. However, we also see that this divergence is quite unexceptional and that HadCRUT + UAH infilling, over the longer period, agrees much stronger with GISS than with HadCRUT.

The main thing that has been going on this decade is actually the divergence in the antarctic, not the arctic. Which is surprising because we have much better coverage in the antarctic 60S-90S both via SSTs and research stations. Also odd, because RSS doesn't do south of 70S but UAH does. Maybe they think it is less accurate?

In the arctic I'm happy to double-check GISS's extrapolations with UAH because have very few stations north of 65N. In the south we have quite a few stations south of 65S, so I am not as eager to double-check GISS using UAH there, given we have reasonable surface data.

[tacoman25]

HadCRUT 60-60 w/ UAH poles is on the other graph.. I didn't include it in that last one because I already had 4 lines and it starts to look cluttered. If you want a specific comparison I can make it pretty quick now that I have all the data in excel.. thinking and making the formulas is what takes time.

I actually did the 1990-1999 base because that allows us to focus on any divergence that has occurred since then (IE over the last decade). I can't make the base a single year because there is a fair amount of noise in single year values. You need at least 10 years I think. And I didn't want to leave it as 1961-1990 because then we would be looking at divergences that have occurred since then. Anyways, it is the year to year changes and trends that are important, which are unaltered by base period selection.

For HadCRUT 60/60 UAH poles vs GISS we do find some divergence if we start in 2000.. you can see that quite clearly if you just look at my 1990-2010 graph of the two and look only at 2000-present (see below). Just look only at the 2000-2010 period and you see there has been some divergence. I actually already pointed out already. However, we also see that this divergence is quite unexceptional and that HadCRUT + UAH infilling, over the longer period, agrees much stronger with GISS than with HadCRUT.

The main thing that has been going on this decade is actually the divergence in the antarctic, not the arctic. Which is surprising because we have much better coverage in the antarctic 60S-90S both via SSTs and research stations.

Ok, I am not trying to be a stickler, but I honestly don't understand this statement; I thought you were in favor of all the sources keeping the same baseline? I don't see how making the baseline 1990-99 is adantageous. Sure, it shows changes compared to the norm in the 1990s, but wouldn't a longer baseline be more representative?

Since Will and I and others have been arguing from the beginning that there has been an increasing divergence with GISS over the past decade (based on normal baselines), all you would have to do to disprove us is make a graph 2000-present, showing trendlines for HadCRU 60/60/UAH poles vs. GISS. That would put all the non-GISS data we reasonably have from the past decade against GISS, and we could clearly see how much, if any, divergence there was.

If there was some divergence, I'm sure it could be argued it could be due to the relatively short time frame. But that graph would be definitive for me, at least.

[skierinvermont]

Ok, I am not trying to be a stickler, but I honestly don't understand this statement; I thought you were in favor of all the sources keeping the same baseline? I don't see how making the baseline 1990-99 is adantageous. Sure, it shows changes compared to the norm in the 1990s, but wouldn't a longer baseline be more representative?

Since Will and I and others have been arguing from the beginning that there has been an increasing divergence with GISS over the past decade (based on normal baselines), all you would have to do to disprove us is make a graph 2000-present, showing trendlines for HadCRU 60/60/UAH poles vs. GISS. That would put all the non-GISS data we reasonably have from the past decade against GISS, and we could clearly see how much, if any, divergence there was.

If there was some divergence, I'm sure it could be argued it could be due to the relatively short time frame. But that graph would be definitive for me, at least.

You can clearly see from the above graph that there IS a divergence this decade. We agree on this fact.. let's move on to the implications of it.

My argument is that the divergence 2000-present between them is just due to the shortness of the period being examined. When we look at longer periods, the divergence disappears and UAH infilling of the poles corroborates GISS.

There are some subtle differences between GISS and surface data with UAH poles which I want to discuss. I just want to agree first that the fact that UAH infilling of the poles 1990-2010 corroborates GISS, and indicates HadCRUT has been biased too cold by excluding the arctic. Using surface data and UAH for the poles seems like a pretty good way of doing things to me, and unequivocally indicates HadCRUT has been biased too cold by excluding the poles.

If we can agree on that, the subtle differences I notice are

1) UAH arctic warms slightly before GISS arctic. UAH arctic warms a lot before 2001, GISS doesn't warm until 2001-2003. This is partly why GISS is biased cold in 1999,2000. It's also partially why GISS diverges some from UAH-infilling of Had or GISS over the last decade. The arctic warming is slightly on GISS. Maybe this is because of the rapid ice melt that occurred.

2) The long term divergence 1990-2010 between GISS antarctic and UAH antarctic. This is fairly moderate compared to the arctic warming, so it only introduces a small divergence globally.

1990-2010 GISS agrees with UAH in the arctic. This accounts for most of the divergence between HadCRUT and GISS over that period. However, for the Antarctic UAH has diverged cooler than GISS somewhat.

I think it would be relevant to decide which of these two sources are more representative of the surface trends... GISS extrapolations or UAH troposphere.

[tacoman25]

You can clearly see from the above graph that there IS a divergence. We agree on this fact.. let's move on to the implications of it.

My argument is that the divergence 2000-present between them is just due to the shortness of the period being examined. When we look at longer periods, the divergence disappears and UAH infilling of the poles corroborates GISS.

There are some subtle differences between GISS and surface data with UAH poles which I want to discuss. I just want to agree first that the fact that UAH infilling of the poles 1990-2010 corroborates GISS, and indicates HadCRUT has been biased too cold by excluding the arctic.

1990-2010 GISS agrees with UAH in the arctic. This accounts for most of the divergence between HadCRUT and GISS over that period. However, for the Antarctic UAH has diverged cooler than GISS somewhat.

I think it would be relevant to decide which of these two sources are more representative of the surface trends... GISS extrapolations or UAH troposphere.

Can you please just make a trendline graph 2000-present with the sources I mentioned? Then we could all clearly see the divergence we have been talking about. I would create it if I had the data/resources at my disposal. I know Will was going to do the same thing eventually.

You have clearly shown what 1990-2010 UAH/GISS Arctic shows. What is of more interest to everyone else, though, is the last decade and both poles UAH/Hadley 60/60 vs. GISS.

And I still don't understand why you are using a 1990-99 baseline in these graphs. Why not longer?

[BethesdaWX]

From you, this is a compliment.

[/size][/color][/color][/size]

No I did not misunderstand you and I didn't read your post quickly. I read it very slowly, as I always do. You specifically stated that both "HadCRUT and GISS have error over the oceans because of sh*tty (sic) coverage and anoms are adjusted to match land anoms"

This is false on two counts. First, there is not "sh*tty" (sic) coverage because they both use satellite SST data. Second, HadCRUT does not use nearby land anoms.

[/size]

More lies that I have disproven 9 times now. GISS has smaller error bars than UAH. UAH also may have substantial methodological bias, which is why it diverges so much from radiosonde and STAR, especially in the mid-troposphere.

If GISS had diverged from UAH substantially over the last decade, then a divergence would still be visible in my graphs of HadCRUT and GISS with UAH infilling of the poles.[/size][/color][/size][/color][/size][/size][/color][/color][/size]

More skier error....shiiaaat!

..... UAH uses RAOBCORE, RICH, HadAT, more data than any other system. AQUA is the most advanced satellite ever put up. Peer reviewed links below.

Your linked Study has been refuted by Roy Spencer, UAH is more accurate than anything since the installation of AQUA in 2002.

GISS is sh*t compared to UAH with its extrapolations...UAH's +/- 0.05C/decade is a base to other instrument due to its lesser "error".

PROBLEMS WITH STAR

2) UAH uses RAOBCORE & RICH too dude, what are you talking about there? You think STAR is the only one that uses differing methods? Difference is the AQUA on UAH implemented in 2002 has eliminated alot of potential error, and is the biggest step forward in satellite technology we have seen thus far.. I linked the Peer Reviewed Study.

The peoblem is STAR uses unproven and un-verifyable Homogenization methods, ok?

Info on AQUA,link to peer reviewed papers inside this link. http://magicjava.blo...-satellite.html

2) I also explained the issues with STAR, it uses an unproven homogenization method that not only assumes errors in the 1970's/80's, but there is no testable verification for it. AQUA satellite puts UAH worlds ahead of STAR regardless.

From Roy Spencer, John Christy:

The standard deviation of these seasonal differences from the mean ranges from 0.050 °C to 0.081 °C. UAH has the smallest standard deviation of "errors" with progressively larger standard deviations for the HadAT, RSS, RAOBCORE, RICH, and RATPAC, respectively. RATPAC has the largest differences likely due to the lower geographic sampling with only 21 stations, a fact referred to later. In Figure 3 we show the individual decadal trends (least squares regression) for years ending in 2005 to 2009 for TLT and Tsfc for direct comparison of all products.

A problematic issue impacts RAOBCORE and RICH and is related to a warming shift in 1991 of the upper troposphere in the ERA-40 Reanalyses on which the two datasets rely. This was shown in [19] to be likely spurious due to a mishandling of a change in an infrared channel, a diagnosis acknowledged by the ECMWF (see also ECMWF Newsletter No. 119, Spring 2009). The shift also led to a sudden and spurious increase in estimates of (a) tropical rainfall, (200 hPa divergence and © low-level humidity. Since this has direct influence on RAOBCORE and to a lesser extent RICH, we would expect these products to display warmer-than-actual trends especially for TMT (which is shown in [9]

[tacoman25]

Skiier, I just want to say that I do respect you and I appreciate all of the hard work you put into research and making these graphs. I do think you have some biases, but don't we all. We disagree on some issues, and I call stuff like I see it, but that doesn't mean I don't like you.

I feel bad that you've had to make so many graphs, it's just that most of them weren't using the information necessary to give the answers many of us are trying to find out!

[ORH_wxman]

I've started some of the GISS stuff today..

I'm starting in 2002 as the point where divergence begins to really increase as that was the first year it was 0.10C warmer than CRU even though it dipped a bit in '03 and early '04...then it ran away from Hadley again.

As for GISS, their arctic trend from 2002-2010 is +0.95C per decade. Their antarctic trend is +0.21C per decade from 2002-2010...inclusive, meaning 9 years of data. The actual whole globe trend is +0.04C per decade since 2002 which is obviously very near flat. So the warming on GISS is coming nearly exclusively from the polar regions on the whole which is not a surprise and already discussed.

What I'm planning to do is use the trend, that way we don't have to try and homogenize the data with a common baseline. (i.e., weight the GISS trend for the 87% of the sfc area in between the poles and then add in the UAH polar trends weighting each one about 6.5%) That should give us an accurate trend for GISS replacing their polar data with UAH.

[skierinvermont]

Here's '02 to present with trends as Will suggested... GISS runs warmer than Had+UAH poles, or regular Had, GISS+UAH poles falls about halfway between.

This tells me that about half the divergence 2002-present between GISS and HadCRUT is due to GISS warming faster 60-60, and the other half is due to its extrapolations at the poles. These trend of these extrapolations at the poles is not supported by UAH 2002-present.

However, I object to the use of trend analysis over such a short period. GISS looks to have extrapolated too cold early in the period, and too warm late in the period. Which could just be luck due to the shortness of the period analyzed. When we look at longer periods, we see that UAH infilling tends to corroborate GISS over HadCRUT. There's just no way for me to possibly believe that HadCRUT is not biased cold when it doesn't include the arctic warming of nearly 1C/decade. The big picture to me is that HadCRUT 60-60+ UAH infilling generally agrees with GISS.

The corollary to the graph below is that 1990-2001 UAH warmed faster than GISS at the poles, and HadCRUT warmed faster than GISS between 60-60. Basically the second graph is the exact opposite of the first graph. Global GISS runs colder than the other 3. From 1990-2001 GISS diverged from HadCRUT partially because it extrapolated too cold at the poles compared to UAH, and partially because HadCRUT warmed more 60-60.

[skierinvermont]

Skiier, I just want to say that I do respect you and I appreciate all of the hard work you put into research and making these graphs. I do think you have some biases, but don't we all. We disagree on some issues, and I call stuff like I see it, but that doesn't mean I don't like you.

I feel bad that you've had to make so many graphs, it's just that most of them weren't using the information necessary to give the answers many of us are trying to find out!

Thank you... I know I can be defensive sometimes but I definitely appreciate and respect people who are willing to dig into this stuff and try to take an objective look at it, which I know you and Will are. There's a big difference between people who actually try to think things through (regardless of any subconscious biases) and people who have made up their minds beforehand, and I can tell all three of us are definitely thinking here.

[ORH_wxman]

That's an interesting thing showing up on GISS for the more recent trends. I haven't gone any deeper than I already posted, but I suspected the same thing when I first eyeballed it. Hadley was running warmer GISS in between the poles in the 1980s and 1990s and assumed it was the same trend in the recent past 9 years but its not. Though its hard to say how much of the difference is beyond 50 degrees on each side...if a lot of it was, then its still sort of the polar thing going on...more like sub-polar.

9 years is not a long time, so I agree the trend analysis can have a large margin for error, but the point of the exercise (for me anyway) was to find out why GISS had been running away from the other agencies in the past decade. Perhaps its just coincidence or variance and they will come back colder in the next decade, but their temps have been a recent point of criticism whether its completely random or not. It also sort of ties in with the whole "we've been flat lining for a while" debate.

[skierinvermont]

Speaking of flat-lining.. I guess the next thing to do would be to start ENSO correcting GISS and HadCRUT w/ UAH infilling as I did for global GISS

Or even better ... average HadCRUT 60S-60N and GISS 60S-60N, then add in UAH for the poles, then (optionally) correct for ENSO... now THAT would be a decisive measure of surface trends

I'm too lazy to do that now so here is 1997-present which is a naturally occurring ENSO neutral period using a two month lag (I checked, the ENSO trend is a hair negative).

Gives a pretty good idea of what the ENSO-corrected trends are at the surface with UAH at the poles. GISS w/ UAH poles is .11C/decade while HadCRUT with UAH poles is .08C/decade. These values fall between the GISS and HadCRUT values of .14C/decade and .04C/decade, respectively.

[BethesdaWX]

Speaking of flat-lining.. I guess the next thing to do would be to start ENSO correcting GISS and HadCRUT w/ UAH infilling as I did for global GISS

Or even better ... average HadCRUT 60S-60N and GISS 60S-60N, then add in UAH for the poles, then (optionally) correct for ENSO... now THAT would be a decisive measure of surface trends

I'm too lazy to do that now so here is 1997-present which is a naturally occurring ENSO neutral period using a two month lag (I checked, the ENSO trend is a hair negative).

Gives a pretty good idea of what the ENSO-corrected trends are at the surface with UAH at the poles. GISS w/ UAH poles is .11C/decade while HadCRUT with UAH poles is .08C/decade. These values fall between the GISS and HadCRUT values of .14C/decade and .04C/decade, respectively.

1) I believe that graph was updated a year ago (2010) Or 6 months ago, a little bit out-dated.

2) Maybe Try adjusting for ENSO, PDO/AMO, and general SST's as well? Using a UAH & GISS comparison without infilling? ( too see the deviation better). The Arctic is capable of the largest deviations of any region in the world in general since it is surrounded by land at the polar lattitude, so it takes very little to affect anoms in a large way.

[skierinvermont]

2) Maybe Try adjusting for ENSO, PDO/AMO, and general SST's as well? Using a UAH & GISS comparison without infilling? ( too see the deviation better).

I'll get right on that.

I'm aware of when the graph was 'updated'.. If you could direct me to data for the final 2011 annual values let me know and I'll plug em in.

[ORH_wxman]

Speaking of flat-lining.. I guess the next thing to do would be to start ENSO correcting GISS and HadCRUT w/ UAH infilling as I did for global GISS

Or even better ... average HadCRUT 60S-60N and GISS 60S-60N, then add in UAH for the poles, then (optionally) correct for ENSO... now THAT would be a decisive measure of surface trends

I'm too lazy to do that now so here is 1997-present which is a naturally occurring ENSO neutral period using a two month lag (I checked, the ENSO trend is a hair negative).

Gives a pretty good idea of what the ENSO-corrected trends are at the surface with UAH at the poles. GISS w/ UAH poles is .11C/decade while HadCRUT with UAH poles is .08C/decade. These values fall between the GISS and HadCRUT values of .14C/decade and .04C/decade, respectively.

Nice graph...can you do one for roughly the two decades prior? Like 1977-1996 or something.

I'm pretty sure the warming was much greater in that period. That probably wouldn't be enso neutral...since I just picked the dates of 2 decades before. It also coincides with the PDO shift in the late '70s.

edit: I guess 1979 would be the earliest since we are using UAH data in those graphs

[skierinvermont]

Nice graph...can you do one for roughly the two decades prior? Like 1977-1996 or something.

I'm pretty sure the warming was much greater in that period. That probably wouldn't be enso neutral...since I just picked the dates of 2 decades before. It also coincides with the PDO shift in the late '70s.

edit: I guess 1979 would be the earliest since we are using UAH data in those graphs

When I first started all this I only imported data back to 1980 so that's where the graphs start. I did the 1980-1996 like you suggested... in that one we see GISS and HadCRUT largely agree... but when we infill the poles with UAH both of them drop considerably which is interesting because UAH infilling of HadCRUT really bumps it up 1997-present.. but really bumps it down 1980-1997

I also went ahead and just did the whole period 1980-present. Trends are as follows. Pretty interesting stuff... GISS is the warmest but GISS w/ UAH poles is the coldest. HadCRUT w/ UAH poles is warmer than HadCRUT. All of them are within less than .02C/decade though.

1980-present

HadCRUT .16C/decade

GISS .168C/decade

HadCRUT w/ UAH poles .162C/decade

GISS w/ UAH poles .153C/decade

[tacoman25]

That's an interesting thing showing up on GISS for the more recent trends. I haven't gone any deeper than I already posted, but I suspected the same thing when I first eyeballed it. Hadley was running warmer GISS in between the poles in the 1980s and 1990s and assumed it was the same trend in the recent past 9 years but its not. Though its hard to say how much of the difference is beyond 50 degrees on each side...if a lot of it was, then its still sort of the polar thing going on...more like sub-polar.

9 years is not a long time, so I agree the trend analysis can have a large margin for error, but the point of the exercise (for me anyway) was to find out why GISS had been running away from the other agencies in the past decade. Perhaps its just coincidence or variance and they will come back colder in the next decade, but their temps have been a recent point of criticism whether its completely random or not. It also sort of ties in with the whole "we've been flat lining for a while" debate.

Yeah, this is what I suspected. I just didn't think the entire GISS divergence with HadCRU over the past 8-9 years could be due just to the Arctic. The overally NH land temps is also driving it...which would support the idea that some of us have suggested that GISS extrapolations may have trended warmer in places besides the Arctic. Doesn't mean they're wrong, but they are warmer recently than the other major temp sources.

It's true that 9 years is not a long period...but the divergence is more significant over parts of that 9 years than any other point in the record. Talking about GISS vs. Hadley and GISS vs. the satellites. That may be meaningful.

[ORH_wxman]

You have a 0 in front of the values per decade...i.e. .016C per decade...do you mean 0.16C per decade?

[ORH_wxman]

Yeah, this is what I suspected. I just didn't think the entire GISS divergence with HadCRU over the past 8-9 years could be due just to the Arctic. The overally NH land temps is also driving it...which would support the idea that some of us have suggested that GISS extrapolations may have trended warmer in places besides the Arctic. Doesn't mean they're wrong, but they are warmer recently than the other major temp sources.

It's true that 9 years is not a long period...but the divergence is more significant over parts of that 9 years than any other point in the record. Talking about GISS vs. Hadley and GISS vs. the satellites. That may be meaningful.

That's why the questions are being asked here. GISS has certainly separated itself from the pack recently...its still a relatively short period so it may be only temporary, but if they continue to diverge, then more questions will be asked.

I think there's some evidence here that GISS is extrapolating too warm in the recent period, but the period in which we are talking about has some decent error margin.

[skierinvermont]

You have a 0 in front of the values per decade...i.e. .016C per decade...do you mean 0.16C per decade?

Ah yes... getting number fatigue .. fixed

[ORH_wxman]

I think a valid question is why 1997-present has basically half the warming of the previous 2 decades...despite a relatively ENSO neutral landscape.

We can blame 2002-present on a -PDO shift I think...though it really didn't take effect in earnest until 2007. A 9 year trend that is actually negative is a bit weird in the AGW modeling...even with with the PDO shift. Its obviously not a huge sample size, but its not pathetic either and likely not to change in the coming several years.

[tacoman25]

I think a valid question is why 1997-present has basically half the warming of the previous 2 decades...despite a relatively ENSO neutral landscape.

We can blame 2002-present on a -PDO shift I think...though it really didn't take effect in earnest until 2007. A 9 year trend that is actually negative is a bit weird in the AGW modeling...even with with the PDO shift. Its obviously not a huge sample size, but its not pathetic either and likely not to change in the coming several years.

For those who believe natural trends are at least 50% responsible for the warming over the past 60 years, it makes perfect sense. For those who believe CO2 was mostly responsible for that warming (and is becoming increasingly dominant as a forcing, which of course the modeling shows), it does seem weird.

[skierinvermont]

That's why the questions are being asked here. GISS has certainly separated itself from the pack recently...its still a relatively short period so it may be only temporary, but if they continue to diverge, then more questions will be asked.

I think there's some evidence here that GISS is extrapolating too warm in the recent period, but the period in which we are talking about has some decent error margin.

Well one thing is to consider is that when I did the graphs and trends for 1990-present.. we all thought that UAH infilling had magically made the divergence disappear (HadCRUT +UAH agreed with normal GISS). In reality, as I suggested and you probably guessed as well, the divergence was still there 2000-present. When I redid the graphs 2000-present... the divergence appeared to reappear .. (HadCRUT+UAH diverged from normal GISS). In reality chopping off 1990-1999 just made it stand out more.. even though it was sitting in plain site before we didn't "see" it. (Although I did suggest that GISS looked too cold 1999-2001 and too warm in 2007).

In other words, the divergence between HadCRUT+UAH infilling and GISS 2000-present was still right there on my graph 1990-present.. but we couldn't see it because it blurred into the general picture of their agreement 1990-present. Even though normal HadCRUT still disagreed with GISS. If normal HadCRUT agreed with GISS 1990-present, then I would say UAH infilling of the poles doesn't "fix" anything. But UAH infilling of the poles "fixes" the divergence 1990-present. Even though it doesn't "fix" the divergence 2000-present. So UAH infilling of the poles is still a powerful explanation for the divergence to me.

[ORH_wxman]

For those who believe natural trends are at least 50% responsible for the warming over the past 60 years, it makes perfect sense. For those who believe CO2 was mostly responsible for that warming (and is becoming increasingly dominant as a forcing, which of course the modeling shows), it does seem weird.

There's a lot of explanations needed though even if its half natural. TSI doesn't correlate well enough to explain it via solar cycles...though geomagnetic flux and cosmic ray theory seems to be doing better....also the lack of knowledge on extended solar max or extended min doesn't help. There is research in its infancy that there is a secondary (and more significant) temp lag during decadal periods of high solar activity and low activity.

Ocean cycles can explain the dip in 1940s-1970s time range and perhaps the recent flattening of the trend, but it shouldn't matter in the long run since we make up for it in the positive/warm ocean cycles.

We also don't all the feedback mechanisms...esp cloud feedback.

[ORH_wxman]

Well one thing is to consider is that when I did the graphs and trends for 1990-present.. we all thought that UAH infilling had magically made the divergence disappear (HadCRUT +UAH agreed with normal GISS). In reality, as I suggested and you probably guessed as well, the divergence was still there 2000-present. When I redid the graphs 2000-present... the divergence appeared to reappear .. (HadCRUT+UAH diverged from normal GISS). In reality chopping off 1990-1999 just made it stand out more.. even though it was sitting in plain site before we didn't "see" it. (Although I did suggest that GISS looked too cold 1999-2001 and too warm in 2007).

In other words, the divergence between HadCRUT+UAH infilling and GISS 2000-present was still right there on my graph 1990-present.. but we couldn't see it because it blurred into the general picture of their agreement 1990-present. Even though normal HadCRUT still disagreed with GISS. If normal HadCRUT agreed with GISS 1990-present, then I would say UAH infilling of the poles doesn't "fix" anything. But UAH infilling of the poles "fixes" the divergence 1990-present. Even though it doesn't "fix" the divergence 2000-present. So UAH infilling of the poles is still a powerful explanation for the divergence to me.

What you say makes sense to a point...but at some point, we need to see GISS come back to the pack. The argument you present is that this past decade is too small a sample to make a big deal out of it...which might be true. But when does it become a big deal? GISS has been criticized recently for warming the present and cooling the past...and whether those claims are scientifically valid (meaning GISS has a legit reason for it) or not still creates some pressure on them since people are focusing on GISS recent trends.

I'd agree that a decade isn't long enough to jump to any conclusions, but its clear they are diverging right now and at some point we'll need to see them come back to the pack (or the pack catch up to them) to make those questions go away.