AMS State of the Climate Report for 2014

[donsutherland1]

The report, which notes that four data sets showed that the global land and ocean temperature anomaly was the highest on record can be found at:

 

http://www2.ametsoc.org/ams/index.cfm/publications/bulletin-of-the-american-meteorological-society-bams/state-of-the-climate/state-of-climate-in-2014/

 

[donsutherland1]

The recent State of the Climate Report for 2014, which noted that four surface temperature data sets agreed that a new global temperature record had been set, has drawn considerable fire from proponents of an alternative climate view. Perhaps because they are unable to refute the data consensus, which poses a challenge to their hypothesis that observed warming has either stopped or become barely discernible in recent years, they have fairly aggressively attacked the data in recent days.

 

Below is one example:

 

 

The question is whether the two charts, neither of which is a surface temperature chart, or the NCDC data set (NOAA) better explains recent surface temperature trends. A secondary question is whether the RSS and UAH data is flawed e.g., biased low, in this case.

 

There is another data set that gains little notice that measures upper air temperatures: RATPAC. Indeed, NOAA has termed RATPAC "the best source of upper air in-situ measurements for studying global temperature trends above the surface." To provide a reasonable proxy for the lower tropospheric readings (surface is part of the lower troposphere, but not identical to the lower troposphere), I took a slice of RATPAC (surface to 500 mb). I also took the GISS data set.

 

Afterward, decadal trends from January 1998 through June 2015 were calculated. Here's what the data showed:

 

GISS: +0.128°C per decade (R²: 0.47)

NCDC: +0.128°C per decade (R²: 0.49)

RATPAC: +0.155°C per decade (R²: 0.29)

 

That RATPAC shows a slightly larger decadal warming trend is consistent with what one would expect when the water vapor greenhouse effect is considered. In short, that both RSS and UAH have diverged sharply from the surface data sets, not to mention the premier upper-air data set, and show an outcome that is inconsistent with the water vapor greenhouse effect, very strongly suggests that the issue lies with the RSS and UAH data sets, not the NCDC data set. IMO, until the flaws in the outlier data sets (RSS and UAH) are understood and addressed, the outliers should be discounted in a serious discussion of global temperature trends. The flaws are of sufficient magnitude that even as the surface data sets and RATPAC (surface to 500 mb) are running their highest warm anomalies on record during the January-June 2015 period, RSS and UAH are showing an entirely different outcome.

[nflwxman]

Great post, Don.  I've heard NOAA has used RATPAC for global temperature trends above the surface for some time.  I believe the IPCC mentions it in passing during it's last report.  UAH/RSS has generally been avoided by the scientific community for TLT trends for a pretty good reason, IMO.

[donsutherland1]

Great post, Don.  I've heard NOAA has used RATPAC for global temperature trends above the surface for some time.  I believe the IPCC mentions it in passing during it's last report.  UAH/RSS has generally been avoided by the scientific community for TLT trends for a pretty good reason, IMO.

Thanks Nflwxman.

[chrisf97212]

Great post, Don.  I've heard NOAA has used RATPAC for global temperature trends above the surface for some time.  I believe the IPCC mentions it in passing during it's last report.  UAH/RSS has generally been avoided by the scientific community for TLT trends for a pretty good reason, IMO.

 

Is RATPAC the one that uses sondes?

[ORH_wxman]

Is RATPAC the one that uses sondes?

 

Yes, and I wouldn't trust it any more than UAH/RSS. I always question the uncertainty bars portrayed on these datasets. Like when people try to tell us that the uncertainty on GISS was plus or minus a few hundreths. The recent revision of SST data source clearly shows that the uncertainty is far greater than claimed. 2014 was adjusted up nearly a full tenth from the new revision. So clearly the uncertainty on last year's temperature was far greater than the published values. The published values are making the assumption that their data is actually accurate, and in most cases, this is not a good assumption.

 

RATPAC is another dataset that is a complete conglomeration of radiosonde measurements and adjustments to produce another dataset seemingly fit for trend analysis...a claimed uncertainty of roughly 0.03C per decade in the lower troposphere. I think that is hogwash.

 

The more I have read up on temperature datasets the past two years in the literature (I'm talking actually reading through the stuff page by page, and not the fancy press release junk or cliffs notes versions on sites like sketpical science, etc, etc), the more I have realized just how difficult it is to actually measure the earth's temperature and trendlines. The adjustments for land surface temps (that are most defined in the USHCN network...less so in the GHCN network) are pretty sound, and I always roll my eyes at the crowd who try to say it's a conspiracy. However, on the flip side, I have now learned to roll my eyes at claimed uncertainties as more and more adjustments come to light.

 

I still question the Karl et al adjustments to buoys...but I'm sure more will come to light on that as it gets more scrutiny from the Hadley team who looked at the same issue the NCDC team did, and determined the adjustments were not necessary in the HadSST3 dataset.

[StudentOfClimatology]

Great post, Don. I've heard NOAA has used RATPAC for global temperature trends above the surface for some time. I believe the IPCC mentions it in passing during it's last report. UAH/RSS has generally been avoided by the scientific community for TLT trends for a pretty good reason, IMO.

Holy moly, there's no way you work in remote sensing. This is simply not true. The UAH/RSS data has been used in over 400 papers since 2010, while the RATPAC sonde aggregation(s) are not even used in the IPCC report, and are scarcely used in the literature.

Actually, the resolution of the RATPAC sonde aggregation(s) are only ~35% of the UAH/RSS data, and extrapolations cover hundreds if not thousands of miles at times. The reason RATPAC is not mainstream is because it's not a good dataset.

[donsutherland1]

 The reason RATPAC is not mainstream is because it's not a good dataset.

 

That's not correct. NOAA has stated on a number of times that it's the "best" measure of tropospheric temperature trends. One example:

 

The best source of upper air in-situ measurements for studying global temperature trends above the surface is the Radiosonde Atmospheric Temperature Products for Assessing Climate (RATPAC) database.

 

https://www.ncdc.noaa.gov/sotc/upper-air/201113

 

While all data sets have limitations, including the GISS and NCDC sets, the ones that are diverging from surface temperature observations are much more suspect than those that provide reasonably consistent findings.

 

That UAH and RSS show nothing remarkable in the face of a strong and strengthening El Niño that has resulted in all four major surface data sets showing that the January-June 2015 anomaly is well ahead ahead of the annual mark last year also flags those two data sets. RATPAC is also showing that the first six months of 2015 produced the highest January-June anomaly on record.

[ORH_wxman]

 

That UAH and RSS show nothing remarkable in the face of a strong and strengthening El Niño that has resulted in all four major surface data sets showing that the January-June 2015 anomaly is well ahead ahead of the annual mark last year also flags those two data sets. RATPAC is also showing that the first six months of 2015 produced the highest January-June anomaly on record.

 

 

I do not think this is great empirical evidence for disqualifying RSS/UAH. If you look back to the strongest El Nino on record in 1997, here were the temp anomalies for both: 

 

RSS:

 

1996 annual temp anomaly: +0.046

1997 first 6 month anomaly: -0.002

 

UAH:

 

1996 annual temp anomaly: -0.075

1997 first 6 month anomaly: -0.155

 

 

 

Both datasets would go on to show unprecedented warming over the next 12 months in response to the historic El Nino event.

 

 

So while I agree that both datasets are quite questionable, it is not because they have lacked the extreme warming like the surface has shown thus far in 2015. Historically, they have had a lagged response to El Ninos where a notable spike was not present until autumn or even as late as mid-winter in a few cases.

[chrisf97212]

.... Historically, they have had a lagged response to El Ninos where a notable spike was not present until autumn or even as late as mid-winter in a few cases.

 

 

Tropical convection being a large driver of that?

[ORH_wxman]

Tropical convection being a large driver of that?

 

 

Absolutely.

[StudentOfClimatology]

That's not correct. NOAA has stated on a number of times that it's the "best" measure of tropospheric temperature trends. One example:

The best source of upper air in-situ measurements for studying global temperature trends above the surface is the Radiosonde Atmospheric Temperature Products for Assessing Climate (RATPAC) database.

https://www.ncdc.noaa.gov/sotc/upper-air/201113

Actually, NOAA's statement does not hold up to scrutiny in the peer reviewed literature. The most recent quality-control analysis (Mears et al 2012) comparing the aggregated sonde networks to the MSU/AMSU sounding units found that the spatial homogeneity was greater in the satellite data than the sonde networks, and that homogenities are internal to both datasets.

http://onlinelibrary.wiley.com/doi/10.1029/2012JD017710/full

Abstract

[1] Multidecadal-scale changes in atmospheric temperature have been measured by both radiosondes and the satellite-borne microwave sounding unit (MSU). Both measurement systems exhibit substantial time varying biases that need to removed to the extent possible from the raw data before they can be used to assess climate trends. A number of methods have been developed for each measurement system, leading to the creation of several homogenized data sets. In this work, we evaluate the agreement between MSU and homogenized radiosonde data sets on multiyear (predominantly 5-year) time scales and find that MSU data sets are often more similar to each other than to radiosonde data sets and vice versa. Furthermore, on these times scales the differences between MSU data sets are often not larger than published internal uncertainty estimates for the RSS product alone and therefore may not be statistically significant when the internal uncertainty in each data set is taken into account. Given the data limitations it is concluded that using radiosondes to validate multidecadal-scale trends in MSU data, or vice versa, or trying to use such metrics alone to pick a ‘winner’ is an ill-conditioned approach and has limited utility without one or more of additional independent measurements, or methodological, or physical analysis.

[donsutherland1]

I do not think this is great empirical evidence for disqualifying RSS/UAH. If you look back to the strongest El Nino on record in 1997, here were the temp anomalies for both: 

 

RSS:

 

1996 annual temp anomaly: +0.046

1997 first 6 month anomaly: -0.002

 

UAH:

 

1996 annual temp anomaly: -0.075

1997 first 6 month anomaly: -0.155

 

 

 

Both datasets would go on to show unprecedented warming over the next 12 months in response to the historic El Nino event.

 

 

So while I agree that both datasets are quite questionable, it is not because they have lacked the extreme warming like the surface has shown thus far in 2015. Historically, they have had a lagged response to El Ninos where a notable spike was not present until autumn or even as late as mid-winter in a few cases.

Good point regarding the lagged ENSO response.  Even if we see a lagged response this time around, the fairly large divergence in temperature trends (January 1998-June 2015) vs. the surface data sets and RATPAC (surface-500mb layers) raises big questions.

 

Furthermore, I looked up the UAH 6.0 beta (which is current only through March), that version's anomalies were as follows for January-March 2015: +0.24°C, +0.14°C, and +0.12°C. Version 5.6 had respective anomalies of +0.36°C, +0.30°C, and +0.25°C. That's a huge difference. For 2014, the V. 6.0 anomaly was +0.15°C; V. 5.6 had an anomaly of +0.28°C. These enormous differences hint that something is wrong. My guess is that the problem lies with the most recent version.

[ORH_wxman]

Good point regarding the lagged ENSO response.  Even if we see a lagged response this time around, the fairly large divergence in temperature trends (January 1998-June 2015) vs. the surface data sets and RATPAC (surface-500mb layers) raises big questions.

 

Furthermore, I looked up the UAH 6.0 beta (which is current only through March), that version's anomalies were as follows for January-March 2015: +0.24°C, +0.14°C, and +0.12°C. Version 5.6 had respective anomalies of +0.36°C, +0.30°C, and +0.25°C. That's a huge difference. For 2014, the V. 6.0 anomaly was +0.15°C; V. 5.6 had an anomaly of +0.28°C. These enormous differences hint that something is wrong. My guess is that the problem lies with the most recent version.

 

Yeah my posted values for UAH were v5.6....without a full peer review of v6.0, we won't know what particular issues are present in that version.

 

It should be noted that the current version of RSS (v3.3) is a peer reviewed dataset. While it looks to have spurious cooling post-2000 in TLT data, it hasn't been debunked in the literature to my knowledge. Until it has been (along with whatever happens to UAH), it would be a bit hasty in my opinion to completely ignore these datasets. UAH (before v6.0) and RSS both had the luxury of being tested against RAOB data as independent sources as part of their verification process. The RATPAC dataset doesn't have this luxury...it is comparing data to itself trying to adjust for time and location. The MSU datasets of course have other issues to account for such as drift, so it also would be ill-advised by me to say they are much less prone to error...but I do not think it is obvious they are more incorrect than RATPAC. The truth could lie somewhere in the middle as well.

[donsutherland1]

Actually, NOAA's statement does not hold up to scrutiny in the peer reviewed literature. The most recent quality-control analysis (Mears et al 2012) comparing the aggregated sonde networks to the MSU/AMSU sounding units found that the spatial homogeneity was greater in the satellite data than the sonde networks, and that homogenities are internal to both datasets.

http://onlinelibrary.wiley.com/doi/10.1029/2012JD017710/full

 

The paper found that using radiosondes to assess which of the MSU datasets was better was not a good approach. That's entirely a different matter from the one that I've raised: the divergence of RSS and UAH trends from those of the major surface data sets. In this case, the radiosondes provide a picture that is consistent with the surface data sets.

 

If at least one of the atmospheric data sets is largely supporting the measured surface temperature trends, the question as to why the MSU ones aren't is worth raising. If a similar situation existed with the atmospheric data sets supporting one another and one or more surface data sets lending support, then the surface data sets that were outliers would need to be re-evaluated.

 

The former situation exists. Therefore, at least for a starting point, the place to look for the issue is the MSU data sets. Perhaps the microwave data doesn't provide a sufficiently robust estimate of surface temperatures and that issue is now becoming evident in the ongoing decoupling of the data.

 

That Version 6.0 of UAH has been running anomalies more than 40% below those of Version 5.6 since at least January 2014 also suggests that 6.0 may have a larger issue than 5.6. Moreover, there has been no peer-reviewed work on the changes made to come up with version 6.0.

 

In the whole scheme of things, the observed warming is continuing (albeit with some differences as to the exact rate). Notions that global temperatures have experienced a small decline in linear trend since January 1998 are suspect.

[donsutherland1]

Yeah my posted values for UAH were v5.6....without a full peer review of v6.0, we won't know what particular issues are present in that version.

 

It should be noted that the current version of RSS (v3.3) is a peer reviewed dataset. While it looks to have spurious cooling post-2000 in TLT data, it hasn't been debunked in the literature to my knowledge. Until it has been (along with whatever happens to UAH), it would be a bit hasty in my opinion to completely ignore these datasets. UAH (before v6.0) and RSS both had the luxury of being tested against RAOB data as independent sources as part of their verification process. The RATPAC dataset doesn't have this luxury...it is comparing data to itself trying to adjust for time and location. The MSU datasets of course have other issues to account for such as drift, so it also would be ill-advised by me to say they are much less prone to error...but I do not think it is obvious they are more incorrect than RATPAC. The truth could lie somewhere in the middle as well.

I don't think one should toss RSS or UAH, but that a closer look should be taken to understand the divergent trends. That RATPAC provides a picture consistent with the surface data sets suggests that perhaps RSS and UAH might have measurement issues (especially version 6.0 of UAH which is providing much lower anomalies than version 5.6). That matter should be examined with attention to assuring that they do, in fact, provide a reliable proxy for temperature trends and, if necessary, developing or refining adjustments to improve such measures.

 

In the context of this issue, I believe it is not constructive to use these data sets as a measure of proof that the trend in the surface data sets is not valid. Until there is strong confidence in the validity of such measures, one should be cautious. Given the issues I raised, particularly the decoupling of trends, real questions exist.

 

Also, the BEST project provided a recent assessment that adds confidence to the major surface data sets. So, at least IMO, one should begin the inquiry by looking at RSS and UAH.

[ORH_wxman]

I don't think one should toss RSS or UAH, but that a closer look should be taken to understand the divergent trends. That RATPAC provides a picture consistent with the surface data sets suggests that perhaps RSS and UAH might have measurement issues (especially version 6.0 of UAH which is providing much lower anomalies than version 5.6). That matter should be examined with attention to assuring that they do, in fact, provide a reliable proxy for temperature trends and, if necessary, developing or refining adjustments to improve such measures.

 

In the context of this issue, I believe it is not constructive to use these data sets as a measure of proof that the trend in the surface data sets is not valid. Until there is strong confidence in the validity of such measures, one should be cautious. Given the issues I raised, particularly the decoupling of trends, real questions exist.

 

Also, the BEST project provided a recent assessment that adds confidence to the major surface data sets. So, at least IMO, one should begin the inquiry by looking at RSS and UAH.

 

I agree that those who use the satellites to disprove the surface are in major error. I do believe the opposite is also true, though our certainty of the surface measurements is stronger than satellites, at least for the land surface temps. I think the sea surface temperature certainty is vastly over-stated, however...which has been proven to an extent by Karl et al (2015) in how much it diverges from Hadsst3 both in the post-2000 trend and overall trend from the mid/late 1800s. What a lot of hardcore skeptics never noted on the new ERSSTv4 dataset was that it lopped off about 0.15C of total warming from the mid/late 1800s. They focus only on the post-2000 differences. Probably mostly by coincidence, these two datasets agree on the trend in the satellite era from 1979...they just get there in vastly different fashions.

[donsutherland1]

I agree that those who use the satellites to disprove the surface are in major error. I do believe the opposite is also true, though our certainty of the surface measurements is stronger than satellites, at least for the land surface temps. I think the sea surface temperature certainty is vastly over-stated, however...which has been proven to an extent by Karl et al (2015) in how much it diverges from Hadsst3 both in the post-2000 trend and overall trend from the mid/late 1800s. What a lot of hardcore skeptics never noted on the new ERSSTv4 dataset was that it lopped off about 0.15C of total warming from the mid/late 1800s. They focus only on the post-2000 differences. Probably mostly by coincidence, these two datasets agree on the trend in the satellite era from 1979...they just get there in vastly different fashions.

I agree, Will, though the warming is solidly supported even if there is some degree of uncertainty in terms of exact magnitude. I'll be interested in seeing how the ERSSTv4 data set fares when some of the other major science agencies have a chance to thoroughly review it.

[nflwxman]

I'd recommend referring to the IPCC, which is obviously the governing force on AGW.  Check out the report of working group 12 in their assessment in climate models.  They seem to favor ECMWF and MERRA reanalysis, as well as RATPAC, over UAH or RSS.  They explicitly mention that the MSU products are too uncertain and large of a swatch of the atmosphere to use in climate model analysis As they state in their summary, there is clearly some benefit to using several measurement stations in LT temperatures versus a single sensor of UAH/RSS.  Their outlier status continues to get larger as this El Nino forms.

 

http://www.climatechange2013.org/images/uploads/WGIAR5_WGI-12Doc2b_FinalDraft_Chapter09.pdf

 

If it's not clear to many posters here now, it should be in a few years that surface temperature climate models, while poor in a sub-decadal sense, do pretty well in a multi-decadal sense.  There is a growing difference between MSU products and every other temperature dataset.  This needs to be addressed at some point.

 

ORH, uncertainties are a little tricky in that they are actually only determined assuming a static methodology of a particular product.  So you can't actually apply uncertainties published for ERSST3 to ERSST4.  It's the same with Hadcrut3 to Hadcrut4 or UAH 5.6 to UAH v6.0.  A better test of the validity of uncertainty analysis is actually to look at the adjustments GISS or Hadley does internal to a static methodology.  In that sense, you almost never see an adjustment of 0.1C or more.

 

More or less- uncertainties are not being applied to global average temperatures, they are only being applied to the dataset that GISS uses to measure global average temperatures (if that makes sense)

[donsutherland1]

I'd recommend referring to the IPCC, which is obviously the governing force on AGW.  Check out the report of working group 12 in their assessment in climate models.  They seem to favor ECMWF and MERRA reanalysis, as well as RATPAC, over UAH or RSS.  They explicitly mention that the MSU products are too uncertain and large of a swatch of the atmosphere to use in climate model analysis As they state in their summary, there is clearly some benefit to using several measurement stations in LT temperatures versus a single sensor of UAH/RSS.  Their outlier status continues to get larger as this El Nino forms.

 

http://www.climatechange2013.org/images/uploads/WGIAR5_WGI-12Doc2b_FinalDraft_Chapter09.pdf

 

If it's not clear to many posters here now, it should be in a few years that surface temperature climate models, while poor in a sub-decadal sense, do pretty well in a multi-decadal sense.  

 

There is a growing difference between MSU products and every other temperature dataset.  This needs to be addressed at some point.

Many thanks for this link. I agree about the growing differences between the MSU products and other temperature records.

[skierinvermont]

I agree with your concerns Don. One other issue is the other studies using MSU and AMSU data that have arrived at very different conclusions than UAH or RSS, such as STAR and the recent Po Chedley paper that strongly critiqued UAH tropical TMT trends. This further reduces the confidence in UAH/RSS results and strengthens those of GISS/BEST/RATPAC.

 

The fact that the IPCC did not even use MSU products for climate model validation speaks volumes.

[AvantHiatus]

Cloud levels (altitude) have been lowering for years and probably troposphere warming along with it. Something big has changed in the system versus the 90's and perhaps 2000's.

[StudentOfClimatology]

The paper found that using radiosondes to assess which of the MSU datasets was better was not a good approach. That's entirely a different matter from the one that I've raised: the divergence of RSS and UAH trends from those of the major surface data sets. In this case, the radiosondes provide a picture that is consistent with the surface data sets.

That's not really what Mears 2012 is doing. They're discussing how to address both internal and structural uncertainty in the MSU/AMSU sounding units and the sonde networks. They're not comparing UAH to RSS, or anything of the sort.

If at least one of the atmospheric data sets is largely supporting the measured surface temperature trends, the question as to why the MSU ones aren't is worth raising. If a similar situation existed with the atmospheric data sets supporting one another and one or more surface data sets lending support, then the surface data sets that were outliers would need to be re-evaluated.

Well first, the surface datasets and satellite datasets don't measure in the same domain(s). They can diverge significantly from one another without one of them being erroneous.

It appears as if UAHv6.0 and RSSv3.3 are in very good agreement now. That may suggest that the RATPAC data is biased warm, or it may not. We'll see what the literature has to say about UAH soon enough.

That Version 6.0 of UAH has been running anomalies more than 40% below those of Version 5.6 since at least January 2014 also suggests that 6.0 may have a larger issue than 5.6. Moreover, there has been no peer-reviewed work on the changes made to come up with version 6.0.

The UAHv6.0 upgrade *supposedly* corrected a slew of problems in the v5.6 data. Again, we'll see what the literature says but this upgrade did bring UAH into lockstep with RSS, so perhaps it's the RATPAC aggregation that is biased warm.

[StudentOfClimatology]

I agree with your concerns Don. One other issue is the other studies using MSU and AMSU data that have arrived at very different conclusions than UAH or RSS, such as STAR and the recent Po Chedley paper that strongly critiqued UAH tropical TMT trends.

The Chedley paper focused only on the TMT, not the TLT. Any studies that "came to different conclusions" than the UAH/RSS teams are old papers critiquing older versions of these datasets.

This further reduces the confidence in UAH/RSS results and strengthens those of GISS/BEST/RATPAC.

This is hyperbolic nonsense, especially considering that the recent GISS update warmed 2014 by 0.093C, which is well above the stated error potential.

[ORH_wxman]

I agree with your concerns Don. One other issue is the other studies using MSU and AMSU data that have arrived at very different conclusions than UAH or RSS, such as STAR and the recent Po Chedley paper that strongly critiqued UAH tropical TMT trends. This further reduces the confidence in UAH/RSS results and strengthens those of GISS/BEST/RATPAC.

 

The fact that the IPCC did not even use MSU products for climate model validation speaks volumes.

 

 

STAR only measures TMT and above.

[stadiumwave]

What I'm reading is RSS & UAH v.6 don't fit the story line so piss on them.

1. I thought the long term story line is what matters so why so much whining? Isn't that the point that was always made when the pre-adjustment hiatus existed & was made much of?

2. Again....the fill in on ARTIC temps on GISS makes me nervous.

3. Honest question...and it may be dumb, but: Why does the NCEP reanalysis 2m temps look more like UAH & RSS? Am I missing something? What is this, what does it mean, & why does it look different than GISS?

 

[StudentOfClimatology]

The CFS-R was never designed to measure global temperatures..there's no quality control for the recurring changes to it's radiative/albedo schemes.

[stadiumwave]

The CFS-R was never designed to measure global temperatures..there's no quality control for the recurring changes to it's radiative/albedo schemes.

My question is what is it for?

[chubbs]

My question is what is it for?

 

Main purpose appears to be for JB and others  to make spurious cooling claims. We've been over this many times on this forum in the past couple of years. This series has problems in 2010 when there was an upgrade. WxBell is the only site I am aware of that stitches the two CFS versions together to make a long-term record.

[donsutherland1]

After looking at the January 1998-December 2014 period (the timeframe where the new UAH is being touted to show a slight cooling trend, which is at odds with the surface datasets and RATPAC), comparing Versions 5.6 and 6.0 of the UAH, I have to state that I'm even more concerned about the reliability of UAH (V 6.0). I ran the numbers for annual Global anomalies and annual anomalies in the North Polar region for the timeframe.

 

Some highlights:

- In terms of annual Global anomalies, the 2001-2014 period saw V.6.0 come out with colder anomalies than V.5.6.

- The 1998-2014 mean Global anomaly was 0.46σ below the V.5.6 mean anomaly for the same timeframe.

- There was a defined trend in which the annual Global anomalies on V6.0 were growing colder by 0.104°C per decade relative to V.5.6 (R²: 0.877)   

- In terms of annual North Polar anomalies, the entire 1998-2014 period saw V.6.0 come out with colder anomalies than V.5.6

- The 1998-2014 mean North Polar anomaly was 0.83σ below the V.5.6 mean anomaly for the same timeframe.

- There was a defined trend in which the annual North Polar anomalies on V.6.0 were growing colder by 0.172°C per decade relative to V.5.6 (R²:0.657) 

 

By itself, the fact that V.6.0 is definitively colder than V.5.6 is one matter (affects estimates of global temperatures, but not trends). That there is a defined trend in which the data on V.6.0 is growing steadily colder relative to V.5.6--and with a high coefficient of correlation--suggests that V.6.0 may suffer from one or more calculation issues that make V.6.0 unreliable in terms of assessing temperature trends. 

 

IMO, V.6.0 urgently needs peer review, especially if there is an embedded trend in which V.6.0 becomes increasingly cold relative to V.5.6. Given that issue, until there is a thorough vetting of UAH V.6.0, its utility for trend analysis is highly suspect.

 

Below is a chart showing the temperature anomalies (°C) on V.5.6 and V.6.0 (with 4/5 years beginning with the previous global temperature record producing a difference of 0.30°C or more in the North Polar region):