2011 Global Temperatures

[skierinvermont]

But if you look globally AND include the Arctic over the last 8-9 years, an increasing divergence is seen with GISS. It's not just one or the other. That is the issue...which you believe to be a non-issue.

No over the last 8-9 years the divergence between HadCRUT w/ UAH poles and GISS is mostly due to between 60S-60N... which is just luck of the draw since they have never really diverged there before.

[BethesdaWX]

The differences are small compared to the overall trend. And those are selective start and end points. If GISS arctic extrapolations were the problem, then replacing them with UAH would make GISS significantly cooler. It doesn't. Case closed.

???

We're talking about the past Decade only...so DUH, yeah, remove the unrelated timeframes earlier on.

[skierinvermont]

Why should I ever believe a method where a large amount of data looks to be unrealistic?

Take a math class. Nobody has any expectation that the GISS extrapolations will be correct in any given month (except you). However, as papers I've posted show, there is no long term bias. This is basic mathematics.

[ORH_wxman]

GISS really wasn't criticized as much prior to 2005...but its recent divergence is what has made it more prone to criticism in recent years. So its plenty valid to look at what its doing in that time. Will it continue? We don't know for sure. My bet is it does because of the way it ignores SST data and extrapolates, which will be more amplified during an ocean cooling phase.

[BethesdaWX]

Take a math class. Nobody has any expectation that the GISS extrapolations will be correct in any given month. However, as papers I've posted show, there is no long term bias.

You're a joke.

Obviously there is a bias over the past Decade! Straying warm...You have no reason to be defending GISS...its because its warmer, and thats why Its laughable.

When GISS is not used in the IPCC widely... thats where suspicion grows.

[skierinvermont]

Yes, we know about the warm phase agreement. I think GISS will continue to diverge from UAH in the polar regions over the next decade. Its clearly doing it more noticeably since the famous 2005 warmest year ever. I think the likely cause of the continued divergence over the next decade will be its lack of use of SST data wherever any ice forms during the year, and since the oceans are in a cooling phase, it will make the difference more and more glaring.

Why would GISS respond to arctic cooling less than UAH?

[skierinvermont]

You're a joke.

Obviously there is a bias over the past Decade! Straying warm...

Not in the long run:

[BethesdaWX]

Why would GISS respond to arctic cooling less than UAH?

He explained it a few pages back, read.

[nzucker]

The differences are small compared to the overall trend. And those are selective start and end points. If GISS arctic extrapolations were the problem, then replacing them with UAH would make GISS significantly cooler. It doesn't. Case closed.

It doesn't really matter where GISS is wrong...the fact is, it's diverged from all the other sources since 2000, and the maps look to contain faulty data.

Look at Greenland this month: RSS showed almost all of Greenland was cooler than normal in March 2011, with some areas nearly 4C below normal. This corroborates Ryan Maue's GFS anomaly maps. Yet, GISS has every single acre of land in Greenland above average for March 2011, with the northwest area more than 4C above average. Surely, GISS must have some stations in Greenland. If that is the case, then how can the satellites find Greenland to be totally in the freezer while GISS says it's torching?

If I said Central Park was -8F for March 2011 and the NWS said it was +8F for the month, would you find that suspicious? Would you say that someone has to be wrong? This isn't any different, Andrew, it doesn't matter whether it's occurring in the Arctic or elsewhere, the problem is that inaccuracies are happening in GISS, on the whole leading to a warm bias.

[BethesdaWX]

Not in the long run:

But we're not talking about the Long run

IN THE PAST DECADE. Remove the sh*t before the early 2000's.

[tacoman25]

GISS really wasn't criticized as much prior to 2005...but its recent divergence is what has made it more prone to criticism in recent years. So its plenty valid to look at what its doing in that time. Will it continue? We don't know for sure. My bet is it does because of the way it ignores SST data and extrapolates, which will be more amplified during an ocean cooling phase.

We mere mortals don't, but skiier knows for sure. That it won't.

Which I guess is why he feels the need to shout us down and call us stupid when we start looking at possible reasons for the divergence.

[BethesdaWX]

We don't, but skiier knows for sure. That it won't.

Which I guess is why he feels the need to shout us down and call us stupid when we start looking at possible reasons for the divergence.

^^ this ^^

[nzucker]

Take a math class. Nobody has any expectation that the GISS extrapolations will be correct in any given month (except you). However, as papers I've posted show, there is no long term bias. This is basic mathematics.

Dude, there's something wrong when one source shows an area +3C and another shows it -3C. That just makes no sense. And these aren't just extrapolations, there must be stations up there.

And yes, there is a bias in the last few years, and it seems to be coming from GISS since it's the outlier run by a known advocate of extreme AGW.

So don't tell me to take a math class...stop thinking you're smarter than everyone in here. I know about the law of large numbers, I was a straight A student in math...this isn't applicable to that law. And I've explained why many, many times. As usual, once you form an opinion you refuse to change it even in the face of overwhelming evidence to the contrary.

[skierinvermont]

GISS really wasn't criticized as much prior to 2005...but its recent divergence is what has made it more prone to criticism in recent years. So its plenty valid to look at what its doing in that time. Will it continue? We don't know for sure. My bet is it does because of the way it ignores SST data and extrapolates, which will be more amplified during an ocean cooling phase.

Well for one thing that coastal extrapolation is a pretty small phenomena.

For another, if the oceans cool, I'd expect coastal cities like Seattle will cool as well, and that cold will then be extrapolated offshore.

Finally, and most importantly, it's arguable which method is closer to the truth. Air temperatures above the ocean are rarely the same as the SSTs. It's common to have warm air 20F+ warmer than the SSTs moving offshore over the water. So I would argue which method is better.. using SSTs or using nearby coastal stations.

[tacoman25]

No over the last 8-9 years the divergence between HadCRUT w/ UAH poles and GISS is mostly due to between 60S-60N... which is just luck of the draw since they have never really diverged there before.

Except that GISS is also diverging with UAH at the poles.

Look at the thread I bumped. Might ring some bells. I'd hate to go through this whole damn discussion again.

[ORH_wxman]

Why would GISS respond to arctic cooling less than UAH?

They don't use SST data any area that ice forms...even if its just briefly during the year. This leads them to use some land temperature and extrapolate the temperatures out in the ocean there...likely to be warmer than reality. Melting sea ice also cools the air at the surface as it does so, so using some land based temperature in the summer months as extrapolating over an area of sea ice is usually going to result in massive error on the warm side.

[skierinvermont]

Dude, there's something wrong when one source shows an area +3C and another shows it -3C. That just makes no sense. And these aren't just extrapolations, there must be stations up there.

And yes, there is a bias in the last few years, and it seems to be coming from GISS since it's the outlier run by a known advocate of extreme AGW.

So don't tell me to take a math class...stop thinking you're smarter than everyone in here. I know about the law of large numbers, I was a straight A student in math...this isn't applicable to that law. And I've explained why many, many times. As usual, once you form an opinion you refuse to change it even in the face of overwhelming evidence to the contrary.

The law of large numbers is absolutely applicable. There is no reason to think that GISS will more frequently extrapolate too warm than too cold.

[BethesdaWX]

Well for one thing that coastal extrapolation is a pretty small phenomena.

For another, if the oceans cool, I'd expect coastal cities like Seattle will cool as well, and that cold will then be extrapolated offshore.

Finally, and most importantly, it's arguable which method is closer to the truth. Air temperatures above the ocean are rarely the same as the SSTs. It's common to have warm air 20F+ warmer than the SSTs moving offshore over the water. So I would argue which method is better.. using SSTs or using nearby coastal stations.

Coastal stations to extrapolate over the Vast oceans? Is that what you are saying?

[tacoman25]

Well for one thing that coastal extrapolation is a pretty small phenomena.

For another, if the oceans cool, I'd expect coastal cities like Seattle will cool as well, and that cold will then be extrapolated offshore.

Finally, and most importantly, it's arguable which method is closer to the truth. Air temperatures above the ocean are rarely the same as the SSTs. It's common to have warm air 20F+ warmer than the SSTs moving offshore over the water. So I would argue which method is better.. using SSTs or using nearby coastal stations.

If I totally trusted GISS, I'd mostly trust their extrapolations. In theory, it shouldn't be a big issue. But we know that subtle differences in methodology can add up over time.

And we also can't ignore the elephant in the room: GISS was started and is ran by a man with a definite agenda and a whole lot of his personal reputation riding on the earth continuing to warm, and in fact, warm faster. That alone makes it difficult for me to completely trust GISS extrapolations, because it is a method where a few tweaks could yield significant differences over time.

[BethesdaWX]

The law of large numbers is absolutely applicable. There is no reason to think that GISS will more frequently extrapolate too warm than too cold.

Except for the fact that it has been doing just that, especially in the Arctic, and is the outlier with the worst resolution and reputation.

[skierinvermont]

They don't use SST data any area that ice forms...even if its just briefly during the year. This leads them to use some land temperature and extrapolate the temperatures out in the ocean there...likely to be warmer than reality. Melting sea ice also cools the air at the surface as it does so, so using some land based temperature in the summer months as extrapolating over an area of sea ice is usually going to result in massive error on the warm side.

Actually, if you think about it the edges of the ice pack (the places that melt in spring and freeze in late fall) have probably warmed faster than anywhere else on earth. Instead of being ice covered 6 months a year they are ice covered 4 or 5 months a year and in those two months they don't have ice when they're supposed to, they are likely much much warmer than climo.

You seem to be forgetting that GISS extrapolates anomalies, not temperatures. It's quite possible GISS will extrapolate a negative coastal anomaly across an area of ocean that melted out a month early and is 10F above average.

[tacoman25]

The law of large numbers is absolutely applicable. There is no reason to think that GISS will more frequently extrapolate too warm than too cold.

Except if you objectively looked at maps, you'd find a whole lot more months like March 2011, and very few where GISS clearly extrapolated cooler than other sources. And what do you know....GISS is actually diverging from other sources in recent years! Coincidence??

I challenge you to find a month that proves GISS extrapolates too cold just as they have too warm last month.

[skierinvermont]

We mere mortals don't, but skiier knows for sure. That it won't.

Which I guess is why he feels the need to shout us down and call us stupid when we start looking at possible reasons for the divergence.

No Will's actually making a reasonable argument that looks at intricate differences between the HadCRUT and GISS method.

Zucker is just blanket blaming the process of extrapolation, which is just a failure to understand the math involved. Numerous studies have shown that the process of extrapolation doesn't create a bias. You can form an accurate index of global temperature using as little as 70 stations and extrapolating 1000s of KM between them. The number of stations GISS uses is extraneous.

[ORH_wxman]

Actually, if you think about it the edges of the ice pack (the places that melt in spring and freeze in late fall) have probably warmed faster than anywhere else on earth. Instead of being ice covered 6 months a year they are ice covered 4 or 5 months a year and in those two months they don't have ice when they're supposed to, they are likely much much warmer than climo.

You seem to be forgetting that GISS extrapolates anomalies, not temperatures. It's quite possible GISS will extrapolate a negative coastal anomaly across an area of ocean that melted out a month early and is 10F above average.

Its using land based temperatures which are going to warm even faster than an area of where sea ice melts out...that's the point. Ice pack retreating from an area near land where it was once lasting longer is going to make that land based station sky rocket.

Regardless, this is a silly debate to go over yet again which was already hashed out in the other thread. GISS is diverging recently and that has continued and I do not see the divergence slowing down.

There are plenty of good reasons to be skeptical of the recent GISS trends.

[BethesdaWX]

Zucker is just blanket blaming the process of extrapolation, which is just a failure to understand the math involved. Numerous studies have shown that the process of extrapolation doesn't create a bias. You can form an accurate index of global temperature using as little as 70 stations and extrapolating 1000s of KM between them. The number of stations GISS uses is extraneous.

WTF? ANd what do you think YOU'RE doing?? Poor arguments defending a datasource which has been straying!

Bottom Line: GISS is too warm because it extrapolates too warm,and this has diverged from everyone else in the past decade.

Now you see why the "+/- .04C decadeargument is so laughable!

[skierinvermont]

Except if you objectively looked at maps, you'd find a whole lot more months like March 2011, and very few where GISS clearly extrapolated cooler than other sources. And what do you know....GISS is actually diverging from other sources in recent years! Coincidence??

I challenge you to find a month that proves GISS extrapolates too cold just as they have too warm last month.

In the long run it does not diverge from HadCRUT or from HadCRUT +UAH infilling over the last 15-30 years.

[BethesdaWX]

In the long run it does not diverge from HadCRUT or from HadCRUT +UAH infilling over the last 15-30 years.

But it does diverge from UAH...boht HADCRUT & GISS have huge dataholes where UAH has data.

So, Fill in UAH data for HADCRUT's Africa, Pole, South America, Australia, & Ocean Dataholes.

[skierinvermont]

Its using land based temperatures which are going to warm even faster than an area of where sea ice melts out...that's the point. Ice pack retreating from an area near land where it was once lasting longer is going to make that land based station sky rocket.

Regardless, this is a silly debate to go over yet again which was already hashed out in the other thread. GISS is diverging recently and that has continued and I do not see the divergence slowing down.

There are plenty of good reasons to be skeptical of the recent GISS trends.

Well as long as we all agree that the proper way to grade them going forward is using HadCRUT w/ UAH poles, or to only compare them between 60S and 60N. Comparing to HadCRUT directly doesn't make much sense since HadCRUT leaves the poles blank. And comparing to UAH or RSS doesn't make much sense since they are likely biased too cold and are not measuring the same thing anyways.

And using either of those two metrics, there has never been a divergence that lasted longer than 15 years. And I see no reason that there will be going forward.

Ice retreating from a shore earlier in the year is going to make the coastal station skyrocket, but it's also going to make the air over water skyrocket because there used to be ice there. Or a month later, say by July, the coastal station won't be that anomalous because the ice has usually melted out near the coast, but the air farther offshore will be really anomalous because usually it is frozen.

[tacoman25]

No Will's actually making a reasonable argument that looks at intricate differences between the HadCRUT and GISS method.

Zucker is just blanket blaming the process of extrapolation, which is just a failure to understand the math involved. Numerous studies have shown that the process of extrapolation doesn't create a bias. You can form an accurate index of global temperature using as little as 70 stations and extrapolating 1000s of KM between them. The number of stations GISS uses is extraneous.

Except that in a region that has very few actual stations (many of which have seen major warming) and a lot of water/ice, extrapolation can definitely tend to be too warm.

Like I said, I mostly trust the theory in a mathematical sense. However, given the evidence of unexplainable divergence globally AND at the poles in recent years, along with a clearly biased scientist running GISS, I find it hard to completely trust this form of extrapolation in reality.

[BethesdaWX]

Well as long as we all agree that the proper way to grade them going forward is using HadCRUT w/ UAH poles, or to only compare them between 60S and 60N.

And using either of those two metrics, there has never been a divergence that lasted longer than 15 years. And I see no reason that there will be going forward.

Ice retreating from a shore earlier in the year is going to make the coastal station skyrocket, but it's also going to make the air over water skyrocket because there used to be ice there. Or a month later, say by July, the coastal station won't be that anomalous because the ice has usually melted out near the coast, but the air farther offshore will be really anomalous because usually it is frozen.

But the anoms over the water will be smaller than those on land, and when you extrapolate the land anoms over water...Bam, GISS is too warm.