GISS vs CRU/RSS/UAH

[tacoman25]

Well the whole assumption of this exercise is that UAH is more accurate than GISS extrapolations. It seems quite probable given the satellite data that GISS extrapolated the arctic too cold in 1999 and 2000 and too warm in 2007.

To your second comment. Again.. you are nitpicking minutiae. If the UAH rise were much slower than the GISS rise, we would see the purple and red lines diverge. We don't, we see general agreement between the red and purple lines. And general agreement between the blue and green lines.

I don't think it's minutiae. Think about it: the greatest period of Arctic divergence (though Arctic difference may not effect global trends much) was easily 2006-2009. Just look at your graph. This was also the period of greatest global differences. This tells me that the differences in the Arctic were probably a microcosm of the growing divergences GISS had with other sources during this period. HadCRU lacking Arctic coverage no doubt caused a greater divergence with GISS, but GISS was still overall likely diverging warmer across the globe. Which is why the satellites have followed GISS a little closer than HadCRUT (including more Arctic temps), but still diverged more as the 2000s wore on.

[tacoman25]

Yeah I would do that except I can't find HadCRUT latitude bands. The CRU site doesn't have it. I used GISS because the NASA site gives 64N-90N which allowed me to subtract off most of the arctic (weighted as only 5%).

That sucks. Because all that graph really proves is that the Arctic temps don't really affect the overall GISS trend.

[skierinvermont]

That sucks. Because all that graph really proves is that the Arctic temps don't really affect the overall GISS trend.

Not true.. because when we took out the arctic it agreed with HadCRUT (green and blue lines). When we added it back in using UAH, it agreed with original GISS (purple and red lines).

[tacoman25]

Not true.. because when we took out the arctic it agreed with HadCRUT (green and blue lines). When we added it back in using UAH, it agreed with original GISS (purple and red lines).

Ah, right. What if you plugged in UAH 60S to 60N, and then used GISS Arctic numbers? Even though theoretically we know what it should look like, it would be nice to see that, as well as the HadCRU + UAH.

[skierinvermont]

Think I found it:

http://climexp.knmi....meone@somewhere

HadCRUT 60-60

Yep.. Bingo:

http://climexp.knmi....hadcrut3_tm.dat

OK Now I will take that and add on UAH 60-85N and see how it compares to GISS.

[skierinvermont]

Here it is:

Graph of HadCRUT 60-60 w/ UAH infilling 60S-90S and 60N-85N; normal HadCRUT; normal GISS. Trendlines are shown for all. The GISS and Had+UAH infilling trend lines overlap which makes them hard to see.

As you can see, HadCRUT w/ UAH infilling generally agrees with GISS. GISS extrapolated too cold in 1999 and 2000 but too warm in 2007 and 2008. The trend is exactly the same.

All base periods were adjusted to 1990-1999.

[BethesdaWX]

Are you arguing GISS is not deviating too warm?

UAH vs GISS. This with the Old UAH baseline. Clearly GISS is too warm, its that simple. No need to argue specifics...

[skierinvermont]

Are you arguing GISS is not deviating too warm?

UAH vs GISS. This with the Old UAH baseline. Clearly GISS is too warm, its that simple. No need to argue specifics...

Selective start and end points... here is full 1990-2010. Essentially complete agreement. The divergence is orders of magnitude smaller than the actual trend.

The trend for UAH is .74C/decade... the trend for GISS is .79C/decade. A disagreement of a mere .05C/decade on a trend of nearly .8C/decade. More than an order of magnitude less than the trend itself. It's of interest scientifically, but for the purpose of 'grading' GISS's arctic extrapolations, UAH strongly confirms GISS. Which is why when we take HadCRUT 60-60 and infill with UAH, we get roughly the same thing as GISS.

[BethesdaWX]

Selective start and end points... here is full 1990-2010. Essentially complete agreement. The divergence is orders of magnitude smaller than the actual trend.

The trend for UAH is .74C/decade... the trend for GISS is .79C/decade. A disagreement of a mere .05C/decade on a trend of nearly .8C/decade. More than an order of magnitude less than the trend itself. It's of interest scientifically, but for the purpose of 'grading' GISS's arctic extrapolations, UAH strongly confirms GISS. Which is why when we take HadCRUT 60-60 and infill with UAH, we get roughly the same thing as GISS.

Uh, dude, the divergence started this decade...thats what is under question and you know it. Don't be a Hypocrite. All modeling systems should use the same base, preferrably 1961-1990.

[BethesdaWX]

Umm, What? Slow down and take a deep breath.

I linked you the info friday. Do we need to go into this all over again? UAH isn't HadAT, UAH is a seperate, more complicated system using over 15 different satellites.

1) Your "study" regarding UAH "error" is no longer valid unfortunately, as Roy Spencer/John Christy have gone into detail on the issue. I'd trust their word on their satellite.

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]

I believe we have this covered as well skier?

good to know

[skierinvermont]

It's sort of surprising that some of you act like the fact that the divergence between HadCRUT and GISS is due to the arctic and antarctic extrapolations, and that these extrapolations are by and large correct, is unheard of.

This has been a widely publicized and acknowledged fact for a while now, including in peer-reviewed literature. GISS has diverged primarily due to its extrapolations of the arctic, and these extrapolations are reasonably accurate, as my graphs show. This is why HadCRUT 60-60 + UAH infilling of the arctic and antarctic agrees with GISS.

[BethesdaWX]

It's sort of surprising that some of you act like the divergence between HadCRUT and GISS is due to the arctic and antarctic extrapolations, and that these extrapolations are by and large correct, is unheard of.

This has been a widely publicized and acknowledged fact for a while now, including in peer-reviewed literature. GISS has diverged primarily due to its extrapolations of the arctic, and these extrapolations are reasonably accurate, as my graphs show. This is why HadCRUT 60-60 + UAH infilling of the arctic and antarctic agrees with GISS.

Dude, the Antarctic has been cooling the entire satellite era.

Are you really arguing tha GISS's warm diversion in the past decade is correct, and everyone else is wrong, despite their superiority? Holy Hell, I think you've jumped off the cliff of pure madness. Its a widely accepted fact that GISS is the outlier, and least likely to be correct. Even the IPCC tends to shy away from GISS...Phil Jones, Michael Mann....dude....

Why do you hug GISS? What is it about GISS you like? Might as well Marry it.

New Flash: Much of GISS's error is possibly over the oceans. You obviously don't have widespread station coverage out there, and surface temps will not vary as much either, so adjusting to surface data to land data variations will lead to error as well.

How about you just go with UAH over GISS? UAH trumps GISS all over, GISS is kind of un-needed actually.

[skierinvermont]

Read more, post less. You don't appear to be following the conversation at all. Your snarky comments, which don't address the facts at hand, are not needed.

[BethesdaWX]

Read more, post less. You don't appear to be following the conversation at all. Your snarky comments, which don't address the facts at hand, are not needed.

How about you think more, post less? I'm speaking of GISS/UAH aone, there is really no need to dive into detail on the matter,its quite simple.

GISS's diversion began in the past decade, not in 1990. The issue is the past decade as far as diversion goes...

[tacoman25]

Selective start and end points... here is full 1990-2010. Essentially complete agreement. The divergence is orders of magnitude smaller than the actual trend.

The trend for UAH is .74C/decade... the trend for GISS is .79C/decade. A disagreement of a mere .05C/decade on a trend of nearly .8C/decade. More than an order of magnitude less than the trend itself. It's of interest scientifically, but for the purpose of 'grading' GISS's arctic extrapolations, UAH strongly confirms GISS. Which is why when we take HadCRUT 60-60 and infill with UAH, we get roughly the same thing as GISS.

Even if you disregard the fact that there HAS been increasing divergence with GISS in both the Arctic and global temperatures over the past decade, the main thing this proves is that adding Arctic warming (different degree with UAH and GISS) slightly alters the temperature trend. But it does not explain why there is a divergence BOTH with the Arctic and the globe with GISS.

If you are going to use the satellite data to prove the Arctic temps play a role (and therefore inherently assert their accuracy), you cannot dismiss the global divergence between satellite temps and GISS/surface temps...which cannot be explained by Arctic temperatures. This is still a problem, and if you are advocating the UAH accuracy in the Arctic, it would be contradictory to blame UAH "flaws" for the global divergence.

[tacoman25]

Here it is:

Graph of HadCRUT 60-60 w/ UAH infilling 60S-90S and 60N-85N; normal HadCRUT; normal GISS. Trendlines are shown for all. The GISS and Had+UAH infilling trend lines overlap which makes them hard to see.

As you can see, HadCRUT w/ UAH infilling generally agrees with GISS. GISS extrapolated too cold in 1999 and 2000 but too warm in 2007 and 2008. The trend is exactly the same.

All base periods were adjusted to 1990-1999.

Another question: why do you insist on starting in 1990 with all of your graphs? You are just ignoring what we have all been saying: THE DIVERGENCE STARTED WELL AFTER THAT.

[tacoman25]

That graph is also a bit deceiving since he started it so early...look at the y-axis values...they are pretty large because it goes back to 1990. I'm focusing on the near term within the post 2002 era where divergence has gotten very large and you can even see on that graph how the latter half of this decade has been so starkly different. If you zoom the graph out far enough it can appear to be minor. I'm going to run the numbers hopefully in the next couple days and post actual values in the recent years...which are sure to be quite different.

Nobody has really questioned the Hadley vs GISS values back in the 1990s....its all been recent. That's where the discrepancy lies, at least with most skeptics...when they are plotted on a graph that goes back to 1990 with larger increments on the y-axis, they don't look as impressive...a recent trend of something like 0.16C per decade vs half that will be hard to see on those.

I just don't have time to run the excel stuff right now with winter wx up here...but I will definitely get around to it. I looke dat the UAH antarctic and arctic trends and they almost completely cancel eachother out in the time frame in question, which means that GISS polar trends that are steeply upward do not match and will make a difference....maybe not an obvious one on a larger graph, but since we are talking only about a decade or even slightly less, you wouldn't expect it to show up on those easily.

Yes, see my post made almost the same time as yours above. I've reminded him repeatedly that starting with 1990 ignores the points we are making, but it seems to be falling on deaf ears, as he keeps going back to that year for some reason.

I am also curious if GISS/HadCRU diverged similarly during previous periods of Arctic cooling. If HadCRU missing the Arctic now is causing it to diverge cooler than GISS, than in cooler Arctic times it should have diverged warmer than GISS. Is this the case?

[ORH_wxman]

Yes, see my post made almost the same time as yours above. I've reminded him repeatedly that starting with 1990 ignores the points we are making, but it seems to be falling on deaf ears, as he keeps going back to that year for some reason.

I am also curious if GISS/HadCRU diverged similarly during previous periods of Arctic cooling. If HadCRU missing the Arctic now is causing it to diverge cooler than GISS, than in cooler Arctic times it should have diverged warmer than GISS. Is this the case?

Like I said, I wish I had the time to crunch the numbers in detail tonight or tomorrow, but I may have to wait until the storms threats have passed here. But I will look at them and then post the graphs. Skier has made some nice graphs, but they still leave a ton of questions as to whether the main areas are being acknowledged and also graphing it out to 1990 will make this decade's trends hard to see....we saw a lot of warming in the 1990s so the y-axis is forced to be large.

I'll post actual yearly values when I run my numbers and the method in which they were achieved. Also his graph is only of arctic (the one you quoted). The recent trend too has been cooling in the southern polar regions. I know he included both poles in an earlier graph, but I found that graph to not filter out enough GISS data, and again it started in the 1990s to widen the y-axis which makes it harder to examine the recent trend. Same with the Hadley+UAH graph...back to 1990 which made more recent trends harder to identify with that spike n the 1990s dominating the y-axis.

As for the arctic cooling, I mentioned that in an earlier post with no response, but not sure if he saw it. The Hadley and GISS seemed to go pretty well from the eyeball test in the arctic cooling period of the 1950-1970s. So GISS was not colder than Hadley from what I saw, but I could be wrong. I will say it was just an eyeball test.

[skierinvermont]

Another question: why do you insist on starting in 1990 with all of your graphs? You are just ignoring what we have all been saying: THE DIVERGENCE STARTED WELL AFTER THAT.

If the divergence starts after 1990, then it would still appear in my graph, and it doesn't. If you want to, put a piece of paper on your screen and cover up everything before the date you would like to start in. I could generate more graphs with later start points, but they would be the exact same as my first graph with the beginning chopped off. Starting in 1990 gives one more perspective. I tried starting in 1980 but that made it hard to see what was happening because it gets too cramped. Just use a piece of paper to cover the part of the graph you don't want to look at.

[skierinvermont]

Even if you disregard the fact that there HAS been increasing divergence with GISS in both the Arctic and global temperatures over the past decade, the main thing this proves is that adding Arctic warming (different degree with UAH and GISS) slightly alters the temperature trend. But it does not explain why there is a divergence BOTH with the Arctic and the globe with GISS.

If you are going to use the satellite data to prove the Arctic temps play a role (and therefore inherently assert their accuracy), you cannot dismiss the global divergence between satellite temps and GISS/surface temps...which cannot be explained by Arctic temperatures. This is still a problem, and if you are advocating the UAH accuracy in the Arctic, it would be contradictory to blame UAH "flaws" for the global divergence.

1) There is no divergence 60S-60N between GISS and HadCRUT. They are identical. See below for graph 60S-60N of HadCRUT and GISS.

(You could have figured this out with some logic. If GISS 60S to 60N was diverging from HadCRUT, then HadCRUT 60S to 60N + UAH infilling would still diverge lower than GISS, and it doesn't. The trends are the same. In fact, I believe HadCRUT 60S-60N has warmed a hair more than GISS.)

2) You keep harping on how I can't use UAH infilling of the poles and simultaneously claim that UAH is biased low globally. Stop and think about this. As I have explained several times, there is no inconsistency here. It is completely consistent to believe that UAH TLT is biased low globally by perhaps a few hundredths C/decade, and yet believe that UAH provides us with a reasonable enough approximation for testing GISS arctic extrapolations. Whether the UAH arctic trend is .74C/decade or .77C/decade, it still corroborates the GISS arctic extrapolations of .79C/decade. And when we substitute it in to GISS or HadCRUT for the poles, the end result = global GISS. This would be true regardless of any small errors in UAH. Why do you keep coming back to this point when I've dispelled it several times?

GISS and HadCRUT 60S-60N exactly the same:

[skierinvermont]

I'll post actual yearly values when I run my numbers and the method in which they were achieved. Also his graph is only of arctic (the one you quoted). The recent trend too has been cooling in the southern polar regions.

I did still include both poles. The method is quite straightforward and definitive in its conclusions. I took HadCRUT 60S-60N and then added on UAH polar data. The result quite clearly shows that HadCRUT 60-60 w/ UAH poles is = to GISS. The graph is shown below once more in case you missed it.

The formula used was .85*(HadCRUT 60S-60N) + .075*(UAH arctic) + .075*(UAH antarctic).

but I found that graph to not filter out enough GISS data,

1) As I explained, I filtered out just as much GISS data as I added back in in UAH warming. I took GISS global anomalies, then subtracted 5% of the earth's surface area, then added back in 5% of the earth's surface area. I could subtract out 7%(GISS), but then I would just have to add back in 7% (UAH) and there would be no net change. This occurs because the UAH arctic has the same trend as the GISS arctic. If GISS and UAH have the same trend for the arctic, which they do, then no matter what you do, replacing one for the other will not lead to major differences. The fact that I am was using UAH 60-85 if anything biases my end result (GISS w/ UAH replacement in poles) too low because 60-85N has slightly less warming than 64-90N.

As for the arctic cooling, I mentioned that in an earlier post with no response, but not sure if he saw it. The Hadley and GISS seemed to go pretty well from the eyeball test in the arctic cooling period of the 1950-1970s. So GISS was not colder than Hadley from what I saw, but I could be wrong. I will say it was just an eyeball test.

Since the base period in the graph you were looking at was set to the cold-arctic period of 1961-1990, you will not find GISS running colder than HadCRUT during that period, since by definition both of them will have an average anomaly of zero during the period for which the base period is set. I get what you are saying and agree, because GISS includes the arctic, it should run colder than HadCRUT during cold arctic periods and warmer during warm arctic periods. However, since the base period IS a cold arctic period, what we will end up seeing is that they are the same during the cold base period, but GISS would be warmer in the 40s and in the 90s-2000s. Which is sort of what we observe (GISS did run warmer in the 40s when the arctic was hot, and also in the 90s and 2000s when it got hot again). However, there are a number of big differences in the SST reconstructions that start to become important when we are looking at such long timescales. The SST reconstruction used by GISS (HadISST1) has some notable differences from HadCRUT's (HadSST2). If we go far enough back, HadCRUT gets quite cold because its SST reconstruction shows more warming (which means it starts colder). The SST reconstruction is the primary reason HadCRUT shows more warming over the last century.

[BethesdaWX]

1) There is no divergence 60S-60N between GISS and HadCRUT. They are identical. See below for graph 60S-60N of HadCRUT and GISS.

(You could have figured this out with some logic. If GISS 60S to 60N was diverging from HadCRUT, then HadCRUT 60S to 60N + UAH infilling would still diverge lower than GISS, and it doesn't. The trends are the same. In fact, I believe HadCRUT 60S-60N has warmed a hair more than GISS.)

2) You keep harping on how I can't use UAH infilling of the poles and simultaneously claim that UAH is biased low globally. Stop and think about this. As I have explained several times, there is no inconsistency here. It is completely consistent to believe that UAH TLT is biased low globally by perhaps a few hundredths C/decade, and yet believe that UAH provides us with a reasonable enough approximation for testing GISS arctic extrapolations. Whether the UAH arctic trend is .74C/decade or .77C/decade, it still corroborates the GISS arctic extrapolations of .79C/decade. And when we substitute it in to GISS or HadCRUT for the poles, the end result = global GISS. This would be true regardless of any small errors in UAH. Why do you keep coming back to this point when I've dispelled it several times?

GISS and HadCRUT 60S-60N exactly the same:

Umm, thats kinda the point. 60S/60N is not the question in regards to GISS/HADCRUT divergence...its the poles over the past decade...that is the whole argument. GISS shows false warming over the Antarctic, as UAH, RSS, HADCRUT, HadAT, RAOBCORE, AQUA, etc (seperate satellites)...all have cooling down there...GISS has sh*tty coverage down there as well, no surprise it shows warming. Arctic has warmed about 1C on UAH since 1994, using GISS's warm extrapolations, it begins to diverge this decade.

Don't bring up the O'donnell paper, because its been explained to you what less coverage means, and the maximum error bars on UAH rule out the possibility of an accurate representation.

HADCRUT/GISS likely feature a certain aspect of Error over the oceans, where there is obviously sh*tty coverage, and anoms are adjusted to match land Anoms...that cannot be done since the lower variations of surface anoms over the oceans are part of the global mean.

[skierinvermont]

Umm, thats kinda the point. 60S/60N is not the question in regards to GISS/HADCRUT divergence...its the poles over the past decade...that is the whole argument. GISS shows false warming over the Antarctic, as UAH, RSS, HADCRUT, HadAT, RAOBCORE, AQUA, etc (seperate satellites)...all have cooling down there...GISS has sh*tty coverage down there as well, no surprise it shows warming. Arctic has warmed about 1C on UAH since 1994, using GISS's warm extrapolations, it begins to diverge this decade.

Don't bring up the O'donnell paper, because its been explained to you what less coverage means, and the maximum error bars on UAH rule out the possibility of an accurate representation.

HADCRUT/GISS likely feature a certain aspect of Error over the oceans, where there is obviously sh*tty coverage, and anoms are adjusted to match land Anoms...that cannot be done since the lower variations of surface anoms over the oceans are part of the global mean.

1) You have no understanding of how the GISS and HadCRUT work. The coverage is best over the oceans because they use satellite measurements of SSTs. You just make up facts when you don't know the answer. Instead of making things up, do some research.

2) I am pleased to hear that you have single handedly refuted the O'Donnell paper. Perhaps you should submit to the Journal of Climate. There is more than enough coverage in the Antarctic and the extrapolations there tend to be quite short compared to other parts of the earth.

3) If GISS diverged from UAH in a major way at the poles, then by substituting in UAH for GISS at the poles, there would be a divergence. There isn't. Taking GISS 60-60 or HadCRUT 60-60 and substituting in UAH for the poles gives approximately the same result as GISS global.

[skierinvermont]

One more variation of the comparison. This one shows that GISS 60-60 and HadCRUT 60-60 are the same, but GISS global and GISS 60-60 w/ UAH poles are also the same. Once again proving that substituting UAH poles into GISS or HadCRUT gives the same result as global GISS, thereby corroborating GISS. The primary reason GISS and HadCRUT have diverged, is the rapid arctic warming which GISS correctly approximates, according to UAH.

[tacoman25]

1) There is no divergence 60S-60N between GISS and HadCRUT. They are identical. See below for graph 60S-60N of HadCRUT and GISS.

(You could have figured this out with some logic. If GISS 60S to 60N was diverging from HadCRUT, then HadCRUT 60S to 60N + UAH infilling would still diverge lower than GISS, and it doesn't. The trends are the same. In fact, I believe HadCRUT 60S-60N has warmed a hair more than GISS.)

2) You keep harping on how I can't use UAH infilling of the poles and simultaneously claim that UAH is biased low globally. Stop and think about this. As I have explained several times, there is no inconsistency here. It is completely consistent to believe that UAH TLT is biased low globally by perhaps a few hundredths C/decade, and yet believe that UAH provides us with a reasonable enough approximation for testing GISS arctic extrapolations. Whether the UAH arctic trend is .74C/decade or .77C/decade, it still corroborates the GISS arctic extrapolations of .79C/decade. And when we substitute it in to GISS or HadCRUT for the poles, the end result = global GISS. This would be true regardless of any small errors in UAH. Why do you keep coming back to this point when I've dispelled it several times?

GISS and HadCRUT 60S-60N exactly the same:

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.

[tacoman25]

If the divergence starts after 1990, then it would still appear in my graph, and it doesn't. If you want to, put a piece of paper on your screen and cover up everything before the date you would like to start in. I could generate more graphs with later start points, but they would be the exact same as my first graph with the beginning chopped off. Starting in 1990 gives one more perspective. I tried starting in 1980 but that made it hard to see what was happening because it gets too cramped. Just use a piece of paper to cover the part of the graph you don't want to look at.

Talk to Will then, too. He had the same questions and believes that starting the trend in 1990 gives an improper perpective, because we have been talking about MORE RECENT trends.

[tacoman25]

Since the base period in the graph you were looking at was set to the cold-arctic period of 1961-1990, you will not find GISS running colder than HadCRUT during that period, since by definition both of them will have an average anomaly of zero during the period for which the base period is set. I get what you are saying and agree, because GISS includes the arctic, it should run colder than HadCRUT during cold arctic periods and warmer during warm arctic periods. However, since the base period IS a cold arctic period, what we will end up seeing is that they are the same during the cold base period, but GISS would be warmer in the 40s and in the 90s-2000s. Which is sort of what we observe (GISS did run warmer in the 40s when the arctic was hot, and also in the 90s and 2000s when it got hot again). However, there are a number of big differences in the SST reconstructions that start to become important when we are looking at such long timescales. The SST reconstruction used by GISS (HadISST1) has some notable differences from HadCRUT's (HadSST2). If we go far enough back, HadCRUT gets quite cold because its SST reconstruction shows more warming (which means it starts colder). The SST reconstruction is the primary reason HadCRUT shows more warming over the last century.

Which goes to show that starting point DOES matter in trend lines.

[BethesdaWX]

1) You have no understanding of how the GISS and HadCRUT work. The coverage is best over the oceans because they use satellite measurements of SSTs. You just make up facts when you don't know the answer. Instead of making things up, do some research.

2) I am pleased to hear that you have single handedly refuted the O'Donnell paper. Perhaps you should submit to the Journal of Climate. There is more than enough coverage in the Antarctic and the extrapolations there tend to be quite short compared to other parts of the earth.

3) If GISS diverged from UAH in a major way at the poles, then by substituting in UAH for GISS at the poles, there would be a divergence. There isn't. Taking GISS 60-60 or HadCRUT 60-60 and substituting in UAH for the poles gives approximately the same result as GISS global.

Holy hell, can you read? You're making yourself look very bad right now.

1) Again, they are measuring Ocean SST's...on land they are measuring air temperature. Thats the issue, there is no AIR TEMPERATURE Coverage over the ocean. You read throug my post So freakin fast, you mis-understand most, if not everyhing I say, then post some unrelated crack.

2) No, I never "refuted the paper", I stated UAH is a more accurate tool to use, and based on its smaller window for error, the conclusions in the paper are somewhat spurious to the rest of the data available.

3) IN THE LAST DECADE, GISS has diverged from UAH......that has beenthe point all along! What don;t you nderstand about that? Again, in the last decade.

[skierinvermont]

Holy hell, can you read? You're making yourself look very bad right now.

From you, this is a compliment.

1) Again, they are measuring Ocean SST's...on land they are measuring air temperature. Thats the issue, there is no AIR TEMPERATURE Coverage over the ocean. You read throug my post So freakin fast, you mis-understand most, if not everyhing I say, then post some unrelated crack.

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.

2) No, I never "refuted the paper", I stated UAH is a more accurate tool to use, and based on its smaller window for error, the conclusions in the paper are somewhat spurious to the rest of the data available.

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.

3) IN THE LAST DECADE, GISS has diverged from UAH......that has beenthe point all along! What don;t you nderstand about that? Again, in the last decade.

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.

[BethesdaWX]

From you, this is a compliment.

That hurt me bad.