Confidence in Surface Air Temperature (SAT)

[longview]

A special topic off of the GISS web site answers questions about how they arrive
at the Surface Air Temperature (SAT).
http://data.giss.nasa.gov/gistemp/abs_temp.html
Some of the answers do not inspire confidence.
Example:

Q. What do we mean by daily mean SAT ?
A. Again, there is no universally accepted correct answer. Should we note the temperature every 6 hours and report the mean, should we do it every 2 hours, hourly, have a machine record it every second, or simply take the average of the highest and lowest temperature of the day ? On some days the various methods may lead to drastically different results.

Their answer makes it sound like they do no know how the per station daily mean is gathered.
I from the more data is better camp, but they should at least be consistent on
sample time and quantity.

At the bottom of the page they talk about using anomaly measurements.

Q. What do I do if I need absolute SATs, not anomalies ?
A. In 99.9% of the cases you'll find that anomalies are exactly what you need, not absolute temperatures. In the remaining cases, you have to pick one of the available climatologies and add the anomalies (with respect to the proper base period) to it. For the global mean, the most trusted models produce a value of roughly 14°C, i.e. 57.2°F, but it may easily be anywhere between 56 and 58°F and regionally, let alone locally, the situation is even worse.

The idea that the global mean "may easily be anywhere between 56 and 58°F",
this is between 13.333 and 14.444 C, 1.1degrees C.
So Per GISS the anomaly reference temperature may be off by 1.1 C.
That error is greater than the total measured increase should be disturbing.

 

[Mithrae]

Their answer makes it sound like they do no know how the per station daily mean is gathered.
I from the more data is better camp, but they should at least be consistent on
sample time and quantity.

In the United States, I believe it's two measurements (a 'minimum' and a 'maximum') each day, as explained in Hansen et al 2010 and no doubt other papers. GISS doesn't have control over data gathering in the United States, let alone globally, and automated stations still haven't completely replaced human observers. Irregularities are inevitable, and corrected for as well as possible.

The idea that the global mean "may easily be anywhere between 56 and 58°F",
this is between 13.333 and 14.444 C, 1.1degrees C.
So Per GISS the anomaly reference temperature may be off by 1.1 C.
That error is greater than the total measured increase should be disturbing.

http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
GISS' global annual anomaly for 2010 was 0.66 degrees Celsius; for 1909, the lowest year on their record, it was -0.46. That's 1.12 C.

 

[longview]

http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
GISS' global annual anomaly for 2010 was 0.66 degrees Celsius; for 1909, the lowest year on their record, it was -0.46. That's 1.12 C.

The anomaly is based on

Best estimate for absolute global mean for 1951-1980 is 14.0 deg-C

From your own citation.
The same site says that the base for the anomaly

may easily be anywhere between 56 and 58°F

This is between 13.333 and 14.444 C, this means the "ZERO" on their tables,
could vary down by an anomaly number of -67 or up by 44 C.
The uncertainty in their ZERO,(1.1 C) is greater than the measured increase, since 1880.

 

[Mithrae]

The anomaly is based on
From your own citation.
The same site says that the base for the anomaly
This is between 13.333 and 14.444 C, this means the "ZERO" on their tables,
could vary down by an anomaly number of -67 or up by 44 C.
The uncertainty in their ZERO,(1.1 C) is greater than the measured increase, since 1880.

It's a bit ambiguous I'll grant you, but your reading just doesn't make sense in the end. Ultimately the question is about absolute temperatures, not the zero reference point for anomalies:

Q. What do I do if I need absolute SATs, not anomalies ?
A. In 99.9% of the cases you'll find that anomalies are exactly what you need, not absolute temperatures. In the remaining cases, you have to pick one of the available climatologies and add the anomalies (with respect to the proper base period) to it. For the global mean, the most trusted models produce a value of roughly 14°C, i.e. 57.2°F, but it may easily be anywhere between 56 and 58°F and regionally, let alone locally, the situation is even worse.

Is the bold part answering the question (ie, giving an idea of what the absolute temperatures will look like), or is it explaining how GISS anomalies should be adjusted? There's three pretty good reasons suggesting the former, namely:

> For adjusting anomalies, the FAQ has directed readers to "one of the available climatologies" and to use "the proper base period," not specifically to GISS' data; the page even begins by stating that "The GISTEMP analysis concerns only temperature anomalies, not absolute temperature," suggesting to readers that it may not be the best place to be looking for absolute temperatures

> If it were explaining how to adjust GISS anomalies into absolute temperatures, it would be pretty much useless, as you are suggesting; it's a general rule of hermeneutics that before concluding that the source is useless/incoherent/stupid, we should look for faults or alternatives in our own understanding of it

> Most importantly, it says that "regionally, let alone locally, the situation is even worse." That makes no sense if it's talking about an average base period temperature to add to anomalies, because calculating those for local or regional data is much easier than globally. But it makes perfect sense if it's trying to give an idea of what the final absolute temperatures will be, because local and regional temperatures do vary a lot more than the global historical record has.

There's also the fact that "the global mean... may easily be anywhere between 56 and 58°F" pretty much exactly matches the range of variation in the GISS global historical record - probably not a coincidence, given the above.

But it might be worthwhile suggesting that they correct that ambiguity.

 

[Visbek]

Their answer makes it sound like they do no know how the per station daily mean is gathered.

They know how it's gathered. They're pointing out some of the uncertainties, in language for a non-technical audience.

This paper has more detail: http://pubs.giss.nasa.gov/docs/2010/2010_Hansen_etal_1.pdf

I from the more data is better camp, but they should at least be consistent on sample time and quantity.

They pretty much explain on that page why that demand really doesn't work, and why it's better to focus on anomalies.

So Per GISS the anomaly reference temperature may be off by 1.1 C. That error is greater than the total measured increase should be disturbing.

No, they're saying that specific areas can vary from the global mean by up to 2º F. They're not talking about a 2º F margin of error.

 

[longview]

It's a bit ambiguous I'll grant you, but your reading just doesn't make sense in the end. Ultimately the question is about absolute temperatures, not the zero reference point for anomalies:

Q. What do I do if I need absolute SATs, not anomalies ?
A. In 99.9% of the cases you'll find that anomalies are exactly what you need, not absolute temperatures. In the remaining cases, you have to pick one of the available climatologies and add the anomalies (with respect to the proper base period) to it. For the global mean, the most trusted models produce a value of roughly 14°C, i.e. 57.2°F, but it may easily be anywhere between 56 and 58°F and regionally, let alone locally, the situation is even worse.

Is the bold part answering the question (ie, giving an idea of what the absolute temperatures will look like), or is it explaining how GISS anomalies should be adjusted? There's three pretty good reasons suggesting the former, namely:

> For adjusting anomalies, the FAQ has directed readers to "one of the available climatologies" and to use "the proper base period," not specifically to GISS' data; the page even begins by stating that "The GISTEMP analysis concerns only temperature anomalies, not absolute temperature," suggesting to readers that it may not be the best place to be looking for absolute temperatures

> If it were explaining how to adjust GISS anomalies into absolute temperatures, it would be pretty much useless, as you are suggesting; it's a general rule of hermeneutics that before concluding that the source is useless/incoherent/stupid, we should look for faults or alternatives in our own understanding of it

> Most importantly, it says that "regionally, let alone locally, the situation is even worse." That makes no sense if it's talking about an average base period temperature to add to anomalies, because calculating those for local or regional data is much easier than globally. But it makes perfect sense if it's trying to give an idea of what the final absolute temperatures will be, because local and regional temperatures do vary a lot more than the global historical record has.

There's also the fact that "the global mean... may easily be anywhere between 56 and 58°F" pretty much exactly matches the range of variation in the GISS global historical record - probably not a coincidence, given the above.

But it might be worthwhile suggesting that they correct that ambiguity.

Their statement is subjective, but they are discussing the anomaly referenced to absolute temperature,
which they state is the average between 1950 and 1980.

As to " how to adjust GISS anomalies into absolute temperatures" that is covered in a foot note in the data.

Best estimate for absolute global mean for 1951-1980 is 14.0 deg-C or 57.2 deg-F,
so add that to the temperature change if you want to use an absolute scale
(this note applies to global annual means only, J-D and D-N !)

 

[longview]

No, they're saying that specific areas can vary from the global mean by up to 2º F. They're not talking about a 2º F margin of error.

I am not sure how you arrived at that interpretation from reading,

Q. What do I do if I need absolute SATs, not anomalies ?
A. In 99.9% of the cases you'll find that anomalies are exactly what you need, not absolute temperatures. In the remaining cases, you have to pick one of the available climatologies and add the anomalies (with respect to the proper base period) to it. For the global mean, the most trusted models produce a value of roughly 14°C, i.e. 57.2°F, but it may easily be anywhere between 56 and 58°F [/B]and regionally, let alone locally, the situation is even worse.

The bold part is describing the global mean,
the follow on in red is saying the variables are even worse regionally and locally.
From your own cited document,

For the sake of users who require an
absolute global mean temperature, we have estimated the
1951–1980 global mean surface air temperature as 14°C
with uncertainty several tenths of a degree Celsius.

The word several is a bit subjective, but between 3 and 4 tenths of a degree C.
So out of the .8 C observed the base reference could be off by .3 to .4 C, wow,
And we have people saying May was the hottest May ever by a whopping .06 C.

 

[Mithrae]

The word several is a bit subjective, but between 3 and 4 tenths of a degree C.
So out of the .8 C observed the base reference could be off by .3 to .4 C, wow,
And we have people saying May was the hottest May ever by a whopping .06 C.

For comparison, HadCRUT's 95% confidence interval for recent years looks to be a bit under 0.2 degrees:

Spoiler

Met Office Hadley Centre observations datasets

But note that while the margin of error creates problems for taking a 0.06 degree difference in one month too seriously, it's much less of a problem for long-term trends. That's because while one year might be up to 0.1 degree warmer or 0.1 degree cooler than the best estimate, it becomes progressively less likely that all years in a period were 0.1 degrees warmer than the best estimate. Over longer periods, it's most likely that the real temperatures would fall on both sides of the estimates, bringing the longer-term averages closer to the center (as shown).

So while 1944 might have been as hot as 1999, 1940-1950 certainly wasn't as hot as 1995-2005.

 

[Manc Skipper]

Dont't you love it when deniers come up with these "Gotcha" threads revealing a basic glaring error that every one of those inside the science conspiracy has (presumably purposely) overlooked. It gets boring when the same "revelations" come around again and again to be swatted.

 

[longview]

For comparison, HadCRUT's 95% confidence interval for recent years looks to be a bit under 0.2 degrees:

Spoiler

Met Office Hadley Centre observations datasets

But note that while the margin of error creates problems for taking a 0.06 degree difference in one month too seriously, it's much less of a problem for long-term trends. That's because while one year might be up to 0.1 degree warmer or 0.1 degree cooler than the best estimate, it becomes progressively less likely that all years in a period were 0.1 degrees warmer than the best estimate. Over longer periods, it's most likely that the real temperatures would fall on both sides of the estimates, bringing the longer-term averages closer to the center (as shown).

So while 1944 might have been as hot as 1999, 1940-1950 certainly wasn't as hot as 1995-2005.

Nice diversion, I was discussing the error range with GISS.
While I have no doubt global temperatures are increasing,
I am looking at how things are averaged, and when the increase occurred.
It is possible to increase an average global temperature, without having any new highs.
This would actually be more consistent with a CO2 response, rather than a solar increase.
I am and always have been skeptical of the open loop feedback of CO2 causing additional warming.
The direct response of CO2, for the most part seems a valid observation.
(Not alarming, but valid).

 

[code1211]

They know how it's gathered. They're pointing out some of the uncertainties, in language for a non-technical audience.

This paper has more detail: http://pubs.giss.nasa.gov/docs/2010/2010_Hansen_etal_1.pdf



They pretty much explain on that page why that demand really doesn't work, and why it's better to focus on anomalies.



No, they're saying that specific areas can vary from the global mean by up to 2º F. They're not talking about a 2º F margin of error.

It's interesting that the folks gathering the data say with authority that the variance can be up to 2 degrees from actual and pin their reputations on calculated temperature changes of a tenth of degree or on a hundredth of degree.

Using the two times a day approach at land stations with various methodologies of data collection, assuming that adjustments will need to be made and knowing that further data changes will be made to the overall seems prone to mischief and error.

It's like using a wooden yardstick to measure the length of single cell organisms. Okay, like using thousands of wooden yardsticks to measure thousands of single cell organisms.

Each measurement is so bad that adjustment is needed. On average, though, it just has to be perfect!

 

[code1211]

Dont't you love it when deniers come up with these "Gotcha" threads revealing a basic glaring error that every one of those inside the science conspiracy has (presumably purposely) overlooked. It gets boring when the same "revelations" come around again and again to be swatted.

You seem to have missed the point. Again.

The methodology is being reviewed.

The outcomes are presented as if the data is beyond challenge. This is not the case. The methodology is open to review and any review of it reveals that the outcome achieved measuring the temperature using techniques from the 19th century is suspect. Also that the methodology varies from one station to the next.

As long as everyone understands that the outcome is close to a variance of 2 degrees away from actual, then that's fine and dandy.

It kind of takes the edge off the claim that we warmed by a half a degree, though.

 

[Jack Hays]

[h=2]Practicing the Dark Art of Temperature Trend Adjustment[/h] Posted on July 4, 2014 by Anthony Watts
Did Federal Climate Scientists Fudge Temperature Data to Make It Warmer?
Ronald Bailey of Reason Magazine writes:
The NCDC also notes that all the changes to the record have gone through peer review and have been published in reputable journals. The skeptics, in turn, claim that a pro-warming confirmation bias is widespread among orthodox climate scientists, tainting the peer review process. Via email, Anthony Watts—proprietor of Watts Up With That, a website popular with climate change skeptics—tells me that he does not think that NCDC researchers are intentionally distorting the record.
But he believes that the researchers have likely succumbed to this confirmation bias in their temperature analyses. In other words, he thinks the NCDC’s scientists do not question the results of their adjustment procedures because they report the trend the researches expect to find. Watts wants the center’s algorithms, computer coding, temperature records, and so forth to be checked by researchers outside the climate science establishment.
Clearly, replication by independent researchers would add confidence to the NCDC results. In the meantime, if Heller episode proves nothing else, it is that we can continue to expect confirmation bias to pervade nearly every aspect of the climate change debate.
Read it all here: Did Federal Climate Scientists Fudge Temperature Data to Make It Warmer? - Reason.com

 

[Mithrae]

As long as everyone understands that the outcome is close to a variance of 2 degrees away from actual, then that's fine and dandy.

It kind of takes the edge off the claim that we warmed by a half a degree, though.

I initially referred him to it and Visbek provided the direct link, so eventually Longview specifically quoted the 2010 GISS update published in Reviews of Geophysics stating that their record has an "uncertainty several tenths of a degree Celsius." For comparison I showed that HadCRUT has a 95% confidence interval of less than 0.2 degrees for recent years (up to around 0.4 degrees in the earliest years covered by GISS). I also explained why the long-term trending has a smaller range of uncertainty than individual years.

Spoiler

Met Office Hadley Centre observations datasets

I can understand that even such basic statistics may not be your strong suit. But given such mathematical shortcomings, obviously your "variance of 2 degrees" is not some carefully worked-out conclusion proving that the experts are wrong by a factor of ten. On the contrary, you've just decided to pull some random number out of thin air. But at least you've made Longview happy to think that someone was naive enough to buy into it :lamo

 

[longview]

I initially referred him to it and Visbek provided the direct link, so eventually Longview specifically quoted the 2010 GISS update published in Reviews of Geophysics stating that their record has an "uncertainty several tenths of a degree Celsius." For comparison I showed that HadCRUT has a 95% confidence interval of less than 0.2 degrees for recent years (up to around 0.4 degrees in the earliest years covered by GISS). I also explained why the long-term trending has a smaller range of uncertainty than individual years.

Spoiler

Met Office Hadley Centre observations datasets

I can understand that even such basic statistics may not be your strong suit. But given such mathematical shortcomings, obviously your "variance of 2 degrees" is not some carefully worked-out conclusion proving that the experts are wrong by a factor of ten. On the contrary, you've just decided to pull some random number out of thin air. But at least you've made Longview happy to think that someone was naive enough to buy into it :lamo

So we have the GISS saying the error in their baseline 1951 to 1980 measurement could be up to 1.1 C, maybe worse locally, in their current web site.

Q. What do I do if I need absolute SATs, not anomalies ?
A. In 99.9% of the cases you'll find that anomalies are exactly what you need, not absolute temperatures. In the remaining cases, you have to pick one of the available climatologies and add the anomalies (with respect to the proper base period) to it. For the global mean, the most trusted models produce a value of roughly 14°C, i.e. 57.2°F, but it may easily be anywhere between 56 and 58°F and regionally, let alone locally, the situation is even worse.

And we have a paper from 2010 saying the same error in the baseline is,

several tenths of a degree Celsius

Is the 2010 paper correct, or the current web site? If they had a better reference from a peer reviewed publication, shouldn't they use it in their web site?
The statement that the regional and local baselines could be

even worse

, is also troubling,
all that pesky significant figure stuff.
The baseline is after all an accumulation of all those regional and local stations.
I started this thread to talk about the method of data collection.
We can safely assume each daily temperature cycle at a single weather station is sinusoidal.
We don't know much about the sample rate, but know the greater the sample rate the better the resolution
of the graph.

 

[code1211]

I initially referred him to it and Visbek provided the direct link, so eventually Longview specifically quoted the 2010 GISS update published in Reviews of Geophysics stating that their record has an "uncertainty several tenths of a degree Celsius." For comparison I showed that HadCRUT has a 95% confidence interval of less than 0.2 degrees for recent years (up to around 0.4 degrees in the earliest years covered by GISS). I also explained why the long-term trending has a smaller range of uncertainty than individual years.

Spoiler

Met Office Hadley Centre observations datasets

I can understand that even such basic statistics may not be your strong suit. But given such mathematical shortcomings, obviously your "variance of 2 degrees" is not some carefully worked-out conclusion proving that the experts are wrong by a factor of ten. On the contrary, you've just decided to pull some random number out of thin air. But at least you've made Longview happy to think that someone was naive enough to buy into it :lamo

I quoted a number used in a post on the topic in a prior post.

 

[Visbek]

It's interesting that the folks gathering the data say with authority that the variance can be up to 2 degrees from actual and pin their reputations on calculated temperature changes of a tenth of degree or on a hundredth of degree.

Again, you're misreading it. They are not saying that the amount of overall uncertainty, or the margin of error, is 2º F per day. What they're saying is that a specific area -- say, New Zealand -- could vary from the global mean by up to 2º F.

Each measurement is so bad that adjustment is needed. On average, though, it just has to be perfect!

The actual scientists working in the field are in fact aware of the uncertainties and measurement issues, and the GISS openly states that in order to deal with the limitations of the measurements, they use computer models.

I might add that these are the same figures and data collection measures that show a pause in the increase of surface temperatures. You cannot, in good faith, simultaneously rely on AND refute the exact same data source.

 

[Mithrae]

So we have the GISS saying the error in their baseline 1951 to 1980 measurement could be up to 1.1 C, maybe worse locally, in their current web site.

And we have a paper from 2010 saying the same error in the baseline is,

Is the 2010 paper correct, or the current web site? If they had a better reference from a peer reviewed publication, shouldn't they use it in their web site?

The statement that the regional and local baselines could be even worse , is also troubling,
all that pesky significant figure stuff.

As I've shown you, and as Visbek has shown you, and as Hansen et al 2010 shows you, and as you are implicitly acknowledging by the absurdity inherent in your interpretation, you are misunderstanding the website. That does not refer to the anomaly baseline period. At this point, we can only assume that you are doing this intentionally - same as your neverending 'misunderstanding' of IPCC temperature projections.


########################
########################

Again, you're misreading it. They are not saying that the amount of overall uncertainty, or the margin of error, is 2º F per day. What they're saying is that a specific area -- say, New Zealand -- could vary from the global mean by up to 2º F.

It actually says that regional and local temperatures can vary by even more than that. Average temperatures in Reykjavik are considerably below 56°F, I'm guessing. The 2º F (c. 1.1º C) refers to the range of variation in the global mean over time, which between the lowest and highest years (1909 and 2010) is 1.12 C.
http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt


########################
########################

I quoted a number used in a post on the topic in a prior post.

Oh, you're going to claim that you meant 2º F? Then out of which bodily orifice did you pull the 0.5º F increase in global temperatures you claimed?

"It kind of takes the edge off the claim that we warmed by a half a degree, though."

 

[Jack Hays]

[h=2]GISS is Unique (Now Includes May Data)[/h] Posted on July 5, 2014 by justthefactswuwt
Guest Essay By Werner Brozek, Edited by Just The Facts
In comparing GISS with the other five data sets that I comment on, some of the points I raise below overlap, and others could be added. However, and in no particular order, the following are some things that I have come up with on why GISS is unique. Perhaps you may disagree on some points or you may come up with others.
Image Credit JoNova
Continue reading →eace

 

[Lord of Planar]

[h=2]GISS is Unique (Now Includes May Data)[/h] Posted on July 5, 2014 by justthefactswuwt
Guest Essay By Werner Brozek, Edited by Just The Facts
In comparing GISS with the other five data sets that I comment on, some of the points I raise below overlap, and others could be added. However, and in no particular order, the following are some things that I have come up with on why GISS is unique. Perhaps you may disagree on some points or you may come up with others.
Image Credit JoNova
Continue reading →eace

Must be what they mean by "improving data." This way, the data says what they want it to say. One more fact making it obviously agenda driven.

 

[Jack Hays]

Must be what they mean by "improving data." This way, the data says what they want it to say. One more fact making it obviously agenda driven.

For whatever reason, we need to know more about the man behind the curtain.eace

 

[Lord of Planar]

For whatever reason, we need to know more about the man behind the curtain.eace

Why?

We already know he has no integrity. What else is there in science? No integrity means no trust of data.

 

[Mithrae]

[h=2]GISS is Unique (Now Includes May Data)[/h] Posted on July 5, 2014 by justthefactswuwt
Guest Essay By Werner Brozek, Edited by Just The Facts
In comparing GISS with the other five data sets that I comment on

The other five including
- HadCRUT3, which is obsolete,
- a sea surface data set SST3, which covers only the least responsive 70% of the Earth's surface
- and RSS, which covers only 70S to 82.5N, and in many ways is even more divergent than GISS

Spoiler

Meanwhile he excludes NCDC, which as the third major English surface temperature record stands alongside HadCRUT4 as the much more obvious point of comparison for GISS than the satellite lower troposphere records. For a couple of the worst cases in which the author's omissions and selection of data are dubious...


3. Including May, GISS has the most months in 2014 above the average of its record year of 2010, namely four of the five months. All other data sets have either zero or one or two months in 2014 above the anomaly average for its highest year. See the table for details.
GISS (going to two decimal places) has January, March, April and May above the 2010 average. HadCRUT4 has January missing out by less than 0.04 degrees, and March missing out in the third decimal place, with April and May exceeding the 2010 average. NCDC has January missing out in the third decimal place, with March, April and May exceeding the 2010 average.
https://www.ncdc.noaa.gov/cag/time-series/global/globe/land_ocean/p12/12/1880-2014.csv
Furthermore, UAH reports monthly anomalies against the 30-year average for that month, so the comparison to UAH (and perhaps RSS) is meaningless.
http://nsstc.uah.edu/climate/2014/may2014/may2014GTR.pdf

6. 1998 is ranked 4th which is the lowest of all data sets. Hadcrut4 has it as third and the others as first.
NCDC and HadCRUT4 both rank the hottest calendar years as 2010, 2005 and 1998. GISS has 2010, 2005, 2007 then 1998. Outside of arbitrary calendar years, the UAH 13-month mean (but not 12-month) also shows 2010 exceeding 1998.

Spoiler

Obviously, GISS measures up just fine against HadCRUT4 and NCDC; the differences are primarily a result of methodology, with GISS' global estimates depending more on extrapolation from nearby stations (out to 1200km) while HadCRUT's extrapolations use the average anomaly per latitude band. Personally, I think the GISS approach makes more sense, but the differences are small enough to be negligable. NCDC's approach lies somewhere between the two.
http://www.metoffice.gov.uk/researc...urface-temperature#differences-in-methodology

http://pubs.giss.nasa.gov/abs/ha00510u.html

http://wattsupwiththat.com/2014/04/...wer-troposphere-temperature-anomaly-datasets/

 

[Jack Hays]

The other five including
- HadCRUT3, which is obsolete,
- a sea surface data set SST3, which covers only the least responsive 70% of the Earth's surface
- and RSS, which covers only 70S to 82.5N, and in many ways is even more divergent than GISS

Spoiler

Meanwhile he excludes NCDC, which as the third major English surface temperature record stands alongside HadCRUT4 as the much more obvious point of comparison for GISS than the satellite lower troposphere records. For a couple of the worst cases in which the author's omissions and selection of data are dubious...


3. Including May, GISS has the most months in 2014 above the average of its record year of 2010, namely four of the five months. All other data sets have either zero or one or two months in 2014 above the anomaly average for its highest year. See the table for details.
GISS (going to two decimal places) has January, March, April and May above the 2010 average. HadCRUT4 has January missing out by less than 0.04 degrees, and March missing out in the third decimal place, with April and May exceeding the 2010 average. NCDC has January missing out in the third decimal place, with March, April and May exceeding the 2010 average.
https://www.ncdc.noaa.gov/cag/time-series/global/globe/land_ocean/p12/12/1880-2014.csv
Furthermore, UAH reports monthly anomalies against the 30-year average for that month, so the comparison to UAH (and perhaps RSS) is meaningless.
http://nsstc.uah.edu/climate/2014/may2014/may2014GTR.pdf

6. 1998 is ranked 4th which is the lowest of all data sets. Hadcrut4 has it as third and the others as first.
NCDC and HadCRUT4 both rank the hottest calendar years as 2010, 2005 and 1998. GISS has 2010, 2005, 2007 then 1998. Outside of arbitrary calendar years, the UAH 13-month mean (but not 12-month) also shows 2010 exceeding 1998.

Spoiler

Obviously, GISS measures up just fine against HadCRUT4 and NCDC; the differences are primarily a result of methodology, with GISS' global estimates depending more on extrapolation from nearby stations (out to 1200km) while HadCRUT's extrapolations use the average anomaly per latitude band. Personally, I think the GISS approach makes more sense, but the differences are small enough to be negligable. NCDC's approach lies somewhere between the two.
Global surface temperature - Met Office

http://pubs.giss.nasa.gov/abs/ha00510u.html

On the Differences and Similarities between Global Surface Temperature and Lower Troposphere Temperature Anomaly Datasets | Watts Up With That?

He includes HadCRUT4.

 

[Threegoofs]

Dont't you love it when deniers come up with these "Gotcha" threads revealing a basic glaring error that every one of those inside the science conspiracy has (presumably purposely) overlooked. It gets boring when the same "revelations" come around again and again to be swatted.

Remember. Scientists aren't recording temperatures correctly.

And if you look at temperatures, the earth is cooling.

But those numbers wouldn't be correct if #1 is true, so just scream Climategate! Every once in a while and everybody will forget that this latest denier position is truly nuts.

This reminds me of Woody Allen:

Diner #1 : the food here is terrible!

Diner #2 : I agree! And such small portions!