2015 Global Temperatures

[AvantHiatus]

That's why we still pick up stuff like ocean fog and smoke as snowcover. Satellite measurements are garbage across the board and we could live without them. They are just a secondary aid.

[nflwxman]

Uh, do you know what the reference to "structural uncertainties" actually means? You're looking at a lot more than aggregated height-based trends here.

I had to do an analysis on that paper back in 2013. Is there a reason you put your faith in Mears et al 2011, despite the fact that it found the largest potential TLT error of any published paper since 2006? It's not exactly "mainstream".

 

SoC.  You are a smart guy, but you get stuck on positions that are clearly not correct.  Instead of getting in the micro about terminology and semantics, try understanding the larger picture.  UAH and RSS both contain higher uncertainties on a multidecadal scale than GISS/HADCrut4 or NOAA.  This is for a multitude of reasons discussed above.  Unless you can provide some type of literature that suggests otherwise, I'm not sure what the point of this debate is.

[StudentOfClimatology]

No, the methodology for the adjustments to surface data have not changed much.

There have been a slew of updates to both the homogenization techniques used in the surface data and the aggregating framework. The resulting change in the observed trend has been more significant than that of the satellite data.

The methodology for satellite data has and continues to change significantly.

What do you mean when you say "the methodology for satellite data"? That's a vague, incoherent statement, imo. The "method" has always been to interpolate microwave emissions in the O^2 spectrum between 50-60GHz. There have been "adjustments" made to correct for drift (easy to do) and other factors like cloud water contamination and diurnal noise (harder to do).

As recently as 1997 Christy and Spencer were claiming the earth was cooling.

That's because the UAH trend from 1979-1997 was/is essentially flat. He was also referring to the TLT, not the planet as a whole (oceans/surface were/are warming.

These are just facts borne out in the literature. If interpreting raw satellite data is 'so easy' and 'such basic stuff' how come the methodology and results have been revised so frequently and are so different from one satellite methodology to another? There's no agreement between sources.

The methodology is the same..the homogenization techniques differ. Same goes for the surface data. Whether you call them "adjustments" or "revisions" is 100% irrelevant.

Revisions are changes in methodology. Adjustments/corrections/interpretation of raw data is something completely different. Raw satellite data is completely useless. Adjustments to correct this are not the same thing as revisions. Revisions are when the methodology that's been used is changed or challenged by an outside source and it has been and is still occurring with satellite data.

Then they're not "revisions" because the underlying methodology behind satellite analysis is unchanged. You're thinking of the parameterization & homogenization techniques that follow the collection of the raw data, which is indeed useless.

The "raw" surface station data is also useless..it needs to be gridded, spatially interpolated, homogenized for contamination, and put through various filters before it is representative of reality.

Honestly, you're trying to create a distinction between the surface data and satellite network that doesn't even exist.

[StudentOfClimatology]

SoC. You are a smart guy, but you get stuck on positions that are clearly not correct. Instead of getting in the micro about terminology and semantics, try understanding the larger picture. UAH and RSS both contain higher uncertainties on a multidecadal scale than GISS/HADCrut4 or NOAA. This is for a multitude of reasons discussed above. Unless you can provide some type of literature that suggests otherwise, I'm not sure what the point of this debate is.

Thanks.

I'm always burying my nose in the literature. But sure, I'll play.

Looks like Christy and Spencer already responded to an earlier version of Fu/Chedley et al, regarding the NOAA9 diurnal noise adjustments they've been trying to apply to UAH. After looking into the earlier paper, I don't find it very compelling, to say the least.

http://journals.ametsoc.org/doi/abs/10.1175/JTECH-D-12-00107.1

The original article that was the subject of this comment/reply can be found at http://journals.ametsoc.org/doi/abs/10.1175/JTECH-D-11-00147.1.

Abstract

Po-Chedley and Fu investigated the difference in the magnitude of global temperature trends generated from the Microwave Sounding Unit (MSU) for the midtroposphere (TMT, surface to about 75 hPa) between the University of Alabama in Huntsville (UAH) and Remote Sensing Systems (RSS). Their approach was to examine the magnitude of a noise-reduction coefficient of one short-lived satellite, NOAA-9, which differed from UAH and RSS. Using radiosonde comparisons over a 2-yr period, they calculated an adjustment to the UAH coefficient that, when applied to the UAH data, increased the UAH global TMT trend for 1979–2009 by +0.042 K decade−1, which then happens to agree with RSS’s TMT trend. In studying their analysis, the authors demonstrate 1) the adjustment calculated using radiosondes is inconclusive when errors are accounted for; 2) the adjustment was applied in a manner inconsistent with the UAH satellite merging strategy, creating a larger change than would be generated had the actual UAH methodology been followed; and 3) that trends of a similar product that uses the same UAH coefficient are essentially identical to UAH and RSS. Based on the authors’ previous analysis and additional work here, UAH will continue using the NOAA-9 noise-reduction coefficient, as is, for version 5.4 and the follow-on version 5.5.

The paper posted here yesterday was using NOAA9, a satellite decommissioned long ago, to interpolate a correction to presumed diurnal noise, despite it's own problems. Not to mention the use of radiosonde data as a controlling framework.

Talk about shaky science.

[StudentOfClimatology]

I have a bunch of these cross-dependent analyses on my computer.

http://www.atmos-meas-tech-discuss.net/8/235/2015/amtd-8-235-2015.pdf

Validation of merged MSU4 and AMSU9 temperature climate records with a new 2002–2012 vertically resolved temperature record

A new database of monthly mean zonal mean (5◦ zones) temperature time series spanning 17 pressure levels from 300 to 7 hPa and extending from 2002 to 2012 was created by merging monthly mean time series from two satellite-based mid-infrared

5 spectrometers (ACE-FTS and MIPAS), a microwave sounder (SMR), and from three satellite-based radio occultation experiments (GRACE, CHAMP, and TSX). The primary intended use of this new temperature data set is to validate the merging of the Microwave Sounding Unit channel 4 (MSU4), and Advanced Microwave Sounding Unit channel 9 (AMSU9) temperature time series conducted in previous studies. The six 10 source data sets were merged by removing offsets and trends between the different measurement series. Weighted means were calculated of the six source data sets where the weights were a function of the uncertainty on the original monthly mean data. This new temperature data set of the upper troposphere and stratosphere has been validated by comparing it to RATPAC-A, COSMIC radio occultation data as well as the 15 NCEPCFSR reanalyses. Differences in all three cases were typically < 2 K in the upper troposphere and lower stratosphere, but could reach up to 5 K in the mid-stratosphere. The data across the 17 pressure levels have then been vertically integrated, using the MSU4/AMSU9 weighting function, to provide a deep vertical layer temperature proxy of the merged MSU4+AMSU9 series. Differences between this vertically integrated data 20 set and two different versions of the MSU4+AMSU9 data set – one from Remote Sens- ing Systems and one from the University of Alabama at Huntsville – were examined for discontinuities. No statistically significant discontinuities were found in either of those two data sets suggesting that the transition from the MSU4+AMSU9 data to AMSU9 data only does not introduce any discontinuities in the MSU4+AMSU9 climate data 25 records that might compromise their use in temperature trend analyses.

[The_Global_Warmer]

It's pretty freaking sad that when it comes to surface data sets versus satellites.  Ones belief about how well each one is almost surely aligns with that they want it to show.

 

 

If the sats were warm and the surface sets cooler.  Not that it really matters since they do not measure the same thing and it's not even close.

 

The same folks would be on different sides.

[The_Global_Warmer]

Major ENSO changes the past two weeks.  Take note this is a 5 or 7 day running average.  Not a daily snapshot.  So changes are smoothed for continuity. 

 

 

 

 

Current forecasts show continued favorable conditions for sub-surface building and Eastward progression of warm pool.

 

Slower warming of Enso 1-2-3.

 

The GFS in a weeks time shows an enhancement of the Easterly flow over enso 3-4.  Nothing major but if that materializes it should be enough to slow the progression and growth of the sub-surface warm pool. 

 

 

 

[nflwxman]

Certainly has the feel of a basin wide moderate event in the making.  This further solidifies the likely 2015 global temperature record.

[The_Global_Warmer]

IMO it's pretty much like a 99.9% done deal at this point.

[StudentOfClimatology]

I'm also leaning towards a record, taking my busted forecast last year into account as well. I learned that statistical predictions often trump dynamical ones on seasonal scales..so definitely taking a different approach this go around.

[ORH_wxman]

I've been on board for 2015 as a record breaker for over a year now when at least warm neutral looked like a good bet for this past winter.

 

I was hedging against 2014, but the big SST binge in the north Pacific really drove us home.

 

 

The question to me is whether we double dip an El Nino so that 2016 can make a run or not.

[chubbs]

 I am thinking slightly warmer on surface mainly in first 7 months. ENSO/SST/PDO all warmer than this time last year and will lap a cool Feb. But will have tougher comparisons Aug-Dec. with 2nd half ENSO uncertain. Satellite should warm a little more than surface since satellite was slower to warm last year.

Agree with posts above. Looks like I was a little too conservative for 2015 based on the warmth so far in Jan-Mar and the strengthening El Nino. Now think that 2015 will see a global  temperature increase similar to 2014's which would bring GISS to roughly 0.75. Going over 0.80 on GISS is more likely at this point than not breaking a record.

[Chinook]

Do you think Mt. Colima explosion on Saturday will affect global temperatures in the next year? The news said that it threw volcanic ash 10,000 ft high. Added to the elevation of 12,000ft, then the ash is approximately at 22,000ft. I heard that great volcano explosions in the tropics: Krakatoa, Pinatubo, and an explosion in Mexico in the early 80's, cooled the global temperature briefly, for about 1 year.

 

http://en.wikipedia.org/wiki/Volc%C3%A1n_de_Colima

[skierinvermont]

Honestly, you're trying to create a distinction between the surface data and satellite network that doesn't even exist.

 

 

The AR5 says that there is a difference and that surface data is much less uncertain. Nothing you've said convinces me otherwise.

[skierinvermont]

That's because the UAH trend from 1979-1997 was/is essentially flat. He was also referring to the TLT, not the planet as a whole (oceans/surface were/are warming.

 

 

The UAH TLT trend from 1979-1997 is not flat. It is nearly .1C/decade. 

 

It was flat (actually slightly negative) before major revisions were done to the UAH methodology. Such large revisions in methodology have never occurred to the surface data sets since their inception.

 

Yet again, you are factually incorrect.

[StudentOfClimatology]

The AR5 says that there is a difference and that surface data is much less uncertain. Nothing you've said convinces me otherwise.

"Structural uncertainty" applies to all time resolutions..this is not a pointed allusion to the long term trend in the data, but rather refers to a myriad of phenomenon like height based warming, month to month changes in tropical convective activity/cloud water contamination, etc. You'd know this if you'd actually read the AR5 report.

The difference between the satellite data tend and the surface network trend is actually within both systems' margin of error. You're making a point of a point until it's pointless.

[StudentOfClimatology]

The UAH TLT trend from 1979-1997 is not flat. It is nearly .1C/decade.

It was flat (actually slightly negative) before major revisions were done to the UAH methodology. Such large revisions in methodology have never occurred to the surface data sets since their inception.

Yet again, you are factually incorrect.

Yet again, you're missing the forest through the trees. The revisions to the aggregated surface station dara were/are larger than that of the satellite data since 1979. Plus these revisions were much more frequent.

[skierinvermont]

Yet again, you're missing the forest through the trees. The revisions to the aggregated surface station dara were/are larger than that of the satellite data since 1979. Plus these revisions were much more frequent.

 

This is false. Please cite changes to the GISS or HadCRUT4 method. The most significant I can think of is the revision from HadCRUT3 to HadCRUT4.

[skierinvermont]

"Structural uncertainty" applies to all time resolutions..this is not a pointed allusion to the long term trend in the data, but rather refers to a myriad of phenomenon like height based warming, month to month changes in tropical convective activity/cloud water contamination, etc. You'd know this if you'd actually read the AR5 report.

The difference between the satellite data tend and the surface network trend is actually within both systems' margin of error. You're making a point of a point until it's pointless.

 

This is not what structural uncertainty means. They specifically say that uncertainty refers to the long-term trend of MSU data.They also specifically say that the uncertainty in trend is greater for satellite than for surface data.

 

You are making a fool of yourself. You are trying to escape the obvious.

[StudentOfClimatology]

This is not what structural uncertainty means.

Umm, yes that exactly what "structural uncertainties" means in the AR5 report..I would've failed my CS105 course had this not been the case. "Structural uncertainty" is a multi-domainal term. You can't flaunt falsehoods and expect to get away with it. I know you're smarter than that.

[StudentOfClimatology]

This is false. Please cite changes to the GISS or HadCRUT4 method. The most significant I can think of is the revision from HadCRUT3 to HadCRUT4.

GISS 1998: http://web.archive.org/web/19990220235952/http:/www.giss.nasa.gov/data/gistemp/GLB.Ts.txt

GISS 2014: http://data.giss.nasa.gov/gistemp/graphs_v3/Fig.A2.txt

I have much older data from the early 1990s on my original computer that I'll link can get to it. The trend differentials are even greater since then.

[The_Global_Warmer]

GISS 1998: http://web.archive.org/web/19990220235952/http:/www.giss.nasa.gov/data/gistemp/GLB.Ts.txt

GISS 2014: http://data.giss.nasa.gov/gistemp/graphs_v3/Fig.A2.txt

I have much older data from the early 1990s on my original computer that I'll link can get to it. The trend differentials are even greater since then.

 

 The top links is land only.

[FloridaJohn]

GISS 1998: http://web.archive.org/web/19990220235952/http:/www.giss.nasa.gov/data/gistemp/GLB.Ts.txt

GISS 2014: http://data.giss.nasa.gov/gistemp/graphs_v3/Fig.A2.txt

I have much older data from the early 1990s on my original computer that I'll link can get to it. The trend differentials are even greater since then.

 

I don't think this is what he requested. My impression was he was looking for a description in how the methodology changed (i.e. what was it before, and how did it change. What was done differently?). I would also be interested in reading this.

 

Here was the original request:

 

Please cite changes to the GISS or HadCRUT4 method.

[StudentOfClimatology]

The top links is land only.

I know says "meteorological stations only", but it's actually a global estimate. Here's the paper where it's from/used:

http://onlinelibrary.wiley.com/doi/10.1029/1999JD900835/abstract

[StudentOfClimatology]

I don't think this is what he requested. My impression was he was looking for a description in how the methodology changed (i.e. what was it before, and how did it change. What was done differently?). I would also be interested in reading this.

The "methodologies" behind both the satellite networks and surface station networks are unchanged..their respective homogenization techniques are what has changed, hence the statistical revisions to both.

My point was that the surface station data has undergone statistical revisions as large or even larger than the satellite data since 1979. This is isn't really disputable.

[ORH_wxman]

The "methodologies" behind both the satellite networks and surface station networks are unchanged..their respective homogenization techniques are what has changed, hence the statistical revisions to both.

My point was that the surface station data has undergone statistical revisions as large or even larger than the satellite data since 1979. This is isn't really disputable.

 

The way the satellite accounts for drift was changed quite a bit after 1997.

 

 

The surface changes are almost all due to GHCN data revisions. Revisions such as TOBS, UHI, LiG to MMTS, etc. There's a few changes made unique to the datset such as going from hadcrut3 to hadcrut4, and GISS switching their SST data back in January 2013 from OISST2 to ERSSTv3....but these have only occurred a few times. NCDC also switched their SST dataset, but this was before GISS did so.

[StudentOfClimatology]

The way the satellite accounts for drift was changed quite a bit after 1997.

The surface changes are almost all due to GHCN data revisions. Revisions such as TOBS, UHI, LiG to MMTS, etc. There's a few changes made unique to the datset such as going from hadcrut3 to hadcrut4, and GISS switching their SST data back in January 2013 from OISST2 to ERSSTv3....but these have only occurred a few times. NCDC also switched their SST dataset, but this was before GISS did so.

Of course, and this clarifies my argument.

Skier was trying to call changes to the surface data "adjustments", but using the term "revisions" to characterize the upgrades to the satellite data, when in reality, they're different words with the same underlying meaning, relative to each dataset's unique properties. Both datasets have undergone significant statistical "revisions", to a very similar extent all around, as a result of changes to quality control techniques. The differences between the satellites and surface datasets are within each network's margin of error.

[skierinvermont]

Of course, and this clarifies my argument.

Skier was trying to call changes to the surface data "adjustments", but using the term "revisions" to characterize the upgrades to the satellite data, when in reality, they're different words with the same underlying meaning, relative to each dataset's unique properties. Both datasets have undergone significant statistical "revisions", to a very similar extent all around, as a result of changes to quality control techniques. The differences between the satellites and surface datasets are within each network's margin of error.

 

There is a very clear difference - everybody else here gets it except you. ORH gets it - he was agreeing with me.

 

One is a change in methodology. 

 

The other is not. 

 

End of story.

 

If I took UAH data for the period 1979-1997 published in 1997 it would look nothing liked data for the same period that is published today. If I took GISS data for 1979-1997 published in 1997, it would look almost identical to GISS data for the same period that is published today. This should adequately illustrate the difference to anybody that doesn't have their head up their arse. Again their have been changes to the GISS, HadCRUT, NOAA methods (the switch to HadCRUT4, the switch in ocean data for NOAA - although I think that was a forward looking change not a change to historical data) but the changes are not nearly as large.

 

 

The former means that all past, present, and future results are changed because an inadequacy in the old method is discoverd - exhibit A is the revisions to satellite data in 1997 that were forced upon UAH by the critiques from RSS. Also, the STAR methodology is a major revision to the pre-existing methodology. The recent paper is yet another significant change in methodology. Each version of UAH has used a somewhat different methodology - yielding corresponding different results. What version of UAH are we on now? Version 5Cz14? When is version 6 coming again?

 

The latter means that past results stay the same. Raw data is adjusted in the present according to a consistent methodology that has been little changed for the last 25+ years. Yes there have been some changes to the methodology of GISS and HadCURT - but the changes to methodology have not yielded changes to historical results that are anywhere near the size of changes that have occurred when UAH/RSS/STAR etc. methodology is changed (IE going from a 20 year cooling trend to a 20 year .1C/decade warming trend).

 

 

Whether you call it revisions adjustments whatever it doesn't matter, that is semantics.

 

In one case the methodology and associated results have been significantly revised multiple times (satellite sources) and in the other the methodology and associated results have remained fairly constant (GISS, HadCRUT).

[skierinvermont]

Umm, yes that exactly what "structural uncertainties" means in the AR5 report..I would've failed my CS105 course had this not been the case. "Structural uncertainty" is a multi-domainal term. You can't flaunt falsehoods and expect to get away with it. I know you're smarter than that.

 

You thinking using big words will make you sound smart? Structural uncertainty simply means all uncertainty inherent in the methodology (IE it includes the uncertainty of the method itself, but also uncertainty that comes from choosing one method over another method). 

 

 

The AR5 is very clear that they are talking about the uncertainty of the long-term trend. It uses the words 'trend' specifically. They're not talking about regional or short-term stuff that you claim they are. 

 

 

Again, here are the quotes:

 

"For large scale trends... uncertainties were concluded to be on the order of .1C/decade."

 

Recent studies of MSU and radiosonde data "highlight the degree of uncertainty in the data and derived products."

 

 

 

Here's another statement that shows they are not talking about regional, or short-term uncertainties, they are talking about long-term trend uncertainties, which they find to be very large: 

 

In summary, assessment of the large body of studies comparing various long-term radiosonde and MSU products since AR4 is hampered by data set version changes, and inherent data uncertainties. These factors substantially limit the ability to draw robust and consistent inferences from such studies about the true long-term trends or the value of different data products.

 

They also mention that comparison of products is difficult because the studies that have done so used "data set versions that have since been superseded." This hits at the issue of significant and frequent changes in methodology to these products that I have referred to. While the surface products remain fairly constant, satellite products have changed methodologies so much that a study comparing UAH to RSS in 2005 would be significantly out of date now because UAH and RSS have both changed their methodology since then and revised pre-2005 data. Whereas a comparison between GISS and NOAA published in 2005 would still be valid today because the data has not changed. 

 

 

Again, they are talking about long-term trends:

 

"Despite unanimous agreement on the sign of the trends, substantial disagreement exists among available estimates as to the rate of temperature changes."

[skierinvermont]

"Structural uncertainty" applies to all time resolutions..this is not a pointed allusion to the long term trend in the data, but rather refers to a myriad of phenomenon like height based warming, month to month changes in tropical convective activity/cloud water contamination, etc. You'd know this if you'd actually read the AR5 report.

The difference between the satellite data tend and the surface network trend is actually within both systems' margin of error. You're making a point of a point until it's pointless.

 

As I point out above, they are specifically referring to the long-term trend. They are not referring to month-to-month changes or height based warming. You'd know this if you'd actually read the AR5 report. 

 

Where are your quotes BethesdaBoy?