2014 Global Temperatures

[The_Global_Warmer]

 Have you seen the Nino OHC anomaly recently? Per NOAA, it has plunged ~2 C in just a few months from +1.95 C in late March to -0.15 C in middle July!! Is this reliable? Where did all of that OHC go? Did it contribute to a warming atmosphere? I'm guessing it did to some extent. I realize this is for just a small volume of the upper oceans of the world to be fair. Smaller volumes mean more volatility. I realize that. But this much volatility? Linked here:

 

 

 

That is measured separately from Argo data.  It probably went into the atmosphere.

[tacoman25]

Given Will's points above, it's mind boggling to me how much trust some put in GCMs and sensitivity estimates.

[StudentOfClimatology]

We generally know that a combination of GHGs, Low Volcanic Activity, Internal Variability, and some measurement bias was responsible for the early 20th century warming.

http://onlinelibrary.wiley.com/doi/10.1029/2000JD000028/abstract

http://link.springer.com/article/10.1023%2FA%3A1005307009726

However, one does not have to consider the early 20th century warming when looking at the difference in temperature between postindustrial and present times. Why? Well, internal and solar forcing agents that caused the early 20th century warming are generally not a factor in present times. Therefore, one would have expected the earth to cool back to it's pre-20th century state without contributions from GHGs and related positive feedbacks. GhGs have forced the climate about 0.8C above preindustrial estimates, regardless of the early 20th century warming period.

I agree somewhat.

Re: The 1910-1945 warming...some agree, some don't. Very different forcings have been highlighted in the literature yet and a similar result is generally achieved depending on the chosen parameterizations, but still fall short in explaining the more extreme climate swings over the past 13Kyrs.

The views on the early 20th century warming are all over the place. I certainly wouldn't jump to any conclusions on the causative mechanism(s).

Scott et al: http://m.sciencemag.org/content/290/5499/2133.short

Shiogama et al: http://onlinelibrary.wiley.com/doi/10.1029/2005GL023540/full

Wang et al:

http://link.springer.com/article/10.1007/s11434-007-0463-y

[StudentOfClimatology]

Have you seen the Nino OHC anomaly recently? Per NOAA, it has plunged ~2 C in just a few months from +1.95 C in late March to -0.15 C in middle July!! Is this reliable? Where did all of that OHC go? Did it contribute to a warming atmosphere? I'm guessing it did to some extent. I realize this is for just a small volume of the upper oceans of the world to be fair. Smaller volumes mean more volatility. I realize that. But this much volatility? Linked here:

http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/lanina/enso_evolution-status-fcsts-web.pdf

Sure the ENSO/KW "contributed", in the sense that it was a major systematic heat release, as seen in the radiation data. But the fact that it forced us into the warmest June on record suggests we may be entering the next step-change (warming) in temperatures, like what we saw in 1998.

[Jesse]

Sure the ENSO/KW "contributed", in the sense that it was a major systematic heat release, as seen in the radiation data. But the fact that it forced us into the warmest June on record suggests we may be entering the next step-change (warming) in temperatures, like what we saw in 1998.

 

Warmest June on record based on just one source whose numbers will probably be further adjusted down the road, correct?

[FloridaJohn]

Warmest June on record based on just one source whose numbers will probably be further adjusted down the road, correct?

Do you think these potential adjustments will move it to one of the coldest Junes on record?

[Isotherm]

Do you think these potential adjustments will move it to one of the coldest Junes on record?

 

 

 

The fact of the matter is, until we see a month or year where there's virtually unanimous agreement between the various datasets, it cannot legitimately be called "warmest on record." Consider this: if there's a jury of 10 people, and 9/10 say let the guy go, but one guy says execute him, should the judge sentence the person to death? In the vein of science, if there's 10 different studies, 8 of which say gravity exists, and 2 say it doesn't, which should we go with? So taking it back to climate, when someone says "this month will be the warmest on record", to me, you will come across as a more objective person if you clarify that phrase with, "according to such and such dataset."

[The_Global_Warmer]

The fact of the matter is, until we see a month or year where there's virtually unanimous agreement between the various datasets, it cannot legitimately be called "warmest on record." Consider this: if there's a jury of 10 people, and 9/10 say let the guy go, but one guy says execute him, should the judge sentence the person to death? In the vein of science, if there's 10 different studies, 8 of which say gravity exists, and 2 say it doesn't, which should we go with? So taking it back to climate, when someone says "this month will be the warmest on record", to me, you will come across as a more objective person if you clarify that phrase with, "according to such and such dataset."

 

 

They all measure different areas.

 

Sat temps measure a huge area volume wise versus ground temps.

 

GISS measures the entire planet while Hadcrut and NCDC do not.

 

By that standard GISS would be the only true surface data set with UAH second in terms of lower troposphere.

 

 

I can't see how you can call Floridajohn less objective when you believe in solar miracles that don't exist.  I doubt you would consider your self less objective over your hope in solar.

[ORH_wxman]

They all measure different areas.

 

Sat temps measure a huge area volume wise versus ground temps.

 

GISS measures the entire planet while Hadcrut and NCDC do not.

 

By that standard GISS would be the only true surface data set with UAH second in terms of lower troposphere.

 

 

I can't see how you can call Floridajohn less objective when you believe in solar miracles that don't exist.  I doubt you would consider your self less objective over your hope in solar.

 

 

This is incorrect...GISS does not "measure" the entire planet. It uses kringing to try and estimate temperatures in the data sparse regions...it also has a unique method in which it uses the closest land temps to kringe areas where sea ice is only seasonal rather than using SST data when it's ice free.

 

This is not to be confused with actual measurements. Generally, it is believed kringing should be able to adequately cover the missing data (esp over land), but it is not without some error and if we are talking about what is measured and what isn't, then it is very important to make that distinction.

[tacoman25]

They all measure different areas.

 

Sat temps measure a huge area volume wise versus ground temps.

 

GISS measures the entire planet while Hadcrut and NCDC do not.

 

By that standard GISS would be the only true surface data set with UAH second in terms of lower troposphere.

 

 

I can't see how you can call Floridajohn less objective when you believe in solar miracles that don't exist.  I doubt you would consider your self less objective over your hope in solar.

 

Not true. GISS extrapolates across large areas that some other sources don't, but it's not like this represents a significant data advantage. These are areas where there just isn't much real data.

[The_Global_Warmer]

This is incorrect...GISS does not "measure" the entire planet. It uses kringing to try and estimate temperatures in the data sparse regions...it also has a unique method in which it uses the closest land temps to kringe areas where sea ice is only seasonal rather than using SST data when it's ice free.

 

This is not to be confused with actual measurements. Generally, it is believed kringing should be able to adequately cover the missing data (esp over land), but it is not without some error and if we are talking about what is measured and what isn't, then it is very important to make that distinction.

 

That is true.  But in terms of June of 2014 the difference between GISS and NCDC is that NCDC has no data at the poles.  It is commonly said here that error in the GISS data extrapolation has a warm bias. I have no idea if that is true.  There have been recent papers or at least one skier posted a while back that found we are underestimating warmth in these data sparse regions.

In other months NCDC is colder because of the poles being the drivers of warmth.

 

NCDC has some pretty large coverage gaps where the coolest anomalies were in June.

 

 

 

 

 

[ORH_wxman]

That is true.  But in terms of June of 2014 the difference between GISS and NCDC is that NCDC has no data at the poles.  It is commonly said here that error in the GISS data extrapolation has a warm bias. I have no idea if that is true.  There have been recent papers or at least one skier posted a while back that found we are underestimating warmth in these data sparse regions.

In other months NCDC is colder because of the poles being the drivers of warmth.

 

NCDC has some pretty large coverage gaps where the coolest anomalies were in June.

 

 

 

 

 

 

 

 

 

Yes, the two datasets definitely differ in their method of estimating the global temps...but it isn't because one has additional data the other doesn't have. It is just how they deal with the "no data" regions.

 

I too think that the kringing method has often underestimated warmth in the arctic. But one aspect of the GISS temp method that gets overlooked is that it often completely ignores large portions of the southern ocean...a region that has cooled rapidly in the past decade. GISS doesn't even try to fill this area in on many of their months...unlike the arctic which they seemingly always fill in. So it is not just a one way street on the nitpicking of the method.

 

 

The methods of each dataset can be debated quite a bit, but in the larger picture, it usually doesn't make much of a difference. It might be the difference between a year ranking #1 or #3 or something...but those are trivial stats, not actually statistically significant since the error bars are larger than the 0.01 or 0.02C that those years are seperated by. I mean, even the record warm year changes on a monthly basis on GISS...sometimes its 2010, sometimes its 2005 and sometimes its tied.

[The_Global_Warmer]

Yes, the two datasets definitely differ in their method of estimating the global temps...but it isn't because one has additional data the other doesn't have. It is just how they deal with the "no data" regions.

 

I too think that the kringing method has often underestimated warmth in the arctic. But one aspect of the GISS temp method that gets overlooked is that it often completely ignores large portions of the southern ocean...a region that has cooled rapidly in the past decade. GISS doesn't even try to fill this area in on many of their months...unlike the arctic which they seemingly always fill in. So it is not just a one way street on the nitpicking of the method.

 

 

The methods of each dataset can be debated quite a bit, but in the larger picture, it usually doesn't make much of a difference. It might be the difference between a year ranking #1 or #3 or something...but those are trivial stats, not actually statistically significant since the error bars are larger than the 0.01 or 0.02C that those years are seperated by. I mean, even the record warm year changes on a monthly basis on GISS...sometimes its 2010, sometimes its 2005 and sometimes its tied.

 

You are probably right on both counts in regards to the SH and NH. 

 

GISS has been really changing a lot recently or I just didn't notice it as much before mid to late 2013.

[ORH_wxman]

You are probably right on both counts in regards to the SH and NH. 

 

GISS has been really changing a lot recently or I just didn't notice it as much before mid to late 2013.

 

 

Well they ditched the HADISST/ReynoldsOISST dataset in January 2013 and switched to ERSSTv3. This caused more shuffling than is normal...so perhaps that is why it seems worse recently. It seems since that switch its been a little more noisey.

[StudentOfClimatology]

Does anyone know why GISS decided to go with ERSST? I personally think that was a stupid move.

[FloridaJohn]

The methods of each dataset can be debated quite a bit, but in the larger picture, it usually doesn't make much of a difference. It might be the difference between a year ranking #1 or #3 or something...but those are trivial stats, not actually statistically significant since the error bars are larger than the 0.01 or 0.02C that those years are seperated by.

This is what I was trying to get at with my earlier question. If every year/month is in the top 10, it doesn't matter if it get changed later from #1 to something lower, it will still be near the top of the records. Unless someone thinks that after later reanalysis it will go from #1 to #118 or something.

 

To me it's not important for all the different data sets to record #1 day/month/year/decade/century/whatever at the same time. If they are generally all headed in the same direction at the same time, then I would call them in agreement. What is the practical difference for a timeperiods temperature to rank #1 by one dataset and #2 by another dataset?

[Isotherm]

This is what I was trying to get at with my earlier question. If every year/month is in the top 10, it doesn't matter if it get changed later from #1 to something lower, it will still be near the top of the records. Unless someone thinks that after later reanalysis it will go from #1 to #118 or something.

 

To me it's not important for all the different data sets to record #1 day/month/year/decade/century/whatever at the same time. If they are generally all headed in the same direction at the same time, then I would call them in agreement. What is the practical difference for a timeperiods temperature to rank #1 by one dataset and #2 by another dataset?

 

 

Well, to me, when or if we attain a new warmest year on record by most objective measures, it signals that our global temperature peaks are now moving above the peaks attained in the 1998 --> 2013 period. For a year to legitimately beat 1998 in terms of warmest on record, that would be a fairly newsworthy IMO even though the statistical/numerical significance of it isn't that great. It would signal that we are no longer falling short of our warmest period 16 years ago.

[Isotherm]

They all measure different areas.

 

Sat temps measure a huge area volume wise versus ground temps.

 

GISS measures the entire planet while Hadcrut and NCDC do not.

 

By that standard GISS would be the only true surface data set with UAH second in terms of lower troposphere.

 

 

I can't see how you can call Floridajohn less objective when you believe in solar miracles that don't exist.  I doubt you would consider your self less objective over your hope in solar.

 

 

 

Maybe we should just throw a huge bucket of water on the Sun, put it out, and see how long humanity lasts off the warmth of Co2.

 

 

*that was a joke by the way* (that is, not to be taken seriously)

[The_Global_Warmer]

1998 was a once in a century event.  It was transient.  Before and after it the Earth was much cooler then it and now.

 

2010 on UAH is virtually tied with 1998.  It shows the progression of warmth.

 

When we take out what if's and projections the Earth is currently quite a bit warmer overall.  If the oceans unloaded heat into the LT at the same equivalency and abruptness of 1998 the result would be much warmer.

 

2005 and 2010 were both warmer than 1998 at the surface.  The difference in terms of the LT is that 1998 unloaded so much heat so fast causing a massive spike in LTs. 

 

The base state today is much warmer.  It may not be as warm as IPCC projections but it's warmer. 

 

The levels in the global ssta charts show this very well.  Inspite of the NINO fail so far.  Global ssts are very very warm.  Record warm at times so far.  In 1997/98 getting above .30C in the NH Spring was not possible without the raging super nino.

 

SSTA have already been running at the levels 1998 only achieved with the Super nino.  2014 achieved them without a nino. 

 

It's a not a perfect comparison but the base state of the oceans is somewhere around .2C higher than 1998. 

 

This is why the surface data sets are outpacing the LTs.  On the surface data sets Nina, Nino, and Neutral years are warmer than the previous corresponding year.

 

 

So far 2014 is running at the records of the warmest years period and has spent the first 5 months in negative neutral land.  2014 will be the warmest neutral year on record which 2013 holds now in the surface data. 

 

Even more impressive is that going back to March of 2007 the PDO/ENSO has been negative the super majority of the time.  Every month in 2013 was negative with a yearly average something like -0.35C.  So far 2014 is still roughly -0.35C in the ONI index but is already on pace to destroy 2013. 

 

2014 has caught or is very close to 2013 on UAH which benefited from a huge January around .49C. 2013 finished around .225C on UAH.  So far 2014 is at .242C.  And the warmest is yet to come. 

 

July of 2013 came in at .174C+.  This year I'd bet the farm that July will double that easy.

 

2013 is the warmest non nino year on every dataset and that is with it being almost a weak nina year.  2014 is going to crush 2013 on every data set with or without a nino. 

 

 

When you compare 2014 to 1997 and 2009 it's not even close on either account. 

 

 

 

 

 

[Sundog]

Well, to me, when or if we attain a new warmest year on record by most objective measures, it signals that our global temperature peaks are now moving above the peaks attained in the 1998 --> 2013 period. For a year to legitimately beat 1998 in terms of warmest on record, that would be a fairly newsworthy IMO even though the statistical/numerical significance of it isn't that great. It would signal that we are no longer falling short of our warmest period 16 years ago.

I don't think it's fair to use the 1998 because it was such a massive Nino. If it was neutral that year that peak would not be there and it would probably fall somewhere around the surrounding years. I think next time we have a very strong Nino we will shatter any previous warm records.

[ORH_wxman]

I don't think it's fair to use the 1998 because it was such a massive Nino. If it was neutral that year that peak would not be there and it would probably fall somewhere around the surrounding years. I think next time we have a very strong Nino we will shatter any previous warm records.

 

Agreed...unfortunately at the time (and even into the early 2000s) it was used to promote a huge positive trend in sfc temps since the late 1970s as proof that global warming was accelerating out of control...rather than pointing to the significance of how strong ENSO events can alter global temperatures.

 

Now, those same people hate it when 1998 is pointed to as a near record (or record) warm year that was 16 years ago.

 

 

Balance in attribution is lacking in climate science discussions all too often. That is what the polarization of a science topic will do.

[Sundog]

Agreed...unfortunately at the time (and even into the early 2000s) it was used to promote a huge positive trend in sfc temps since the late 1970s as proof that global warming was accelerating out of control...rather than pointing to the significance of how strong ENSO events can alter global temperatures.

Now, those same people hate it when 1998 is pointed to as a near record (or record) warm year that was 16 years ago.

Balance in attribution is lacking in climate science discussions all too often. That is what the polarization of a science topic will do.

Agreed.

[Isotherm]

I agree that 1998 was a once in 60+ year type of event in terms of magnitude, and the chances of seeing a +ENSO event of similar intensity likely won't be until the return to the +PDO decadal phase 20-25+ years down the road.

 

With that being said, if the global temp peaks coinciding w/ strong nino's continue increasing in a linear fashion, we can tell that the overall base state is warming. The 2009-10 event was a fairly recent, very potent Nino. Would it have beaten 1998 if it were > +2.0c? Possibly. If it did, it's still difficult to isolate causation variables and their corresponding percentage influences. The Atlantic was in a predominately cold phase prior to the 1997-98 event which probably aided in a lower launching point in terms of global temperatures. The largely warm AMO from the late 90s through the 2000s coupled with the +PDO (until 2007) helped create a warmer base state / starting point prior to the 2009 event.

[FloridaJohn]

With that being said, if the global temp peaks coinciding w/ strong nino's continue increasing in a linear fashion, we can tell that the overall base state is warming.

 

You mean like this?

 

[tacoman25]

Personally, I think the best way to gauge meaningful long term warming is what is going on the surface of land masses, in areas that haven't seen massive land changes over the period of record. SSTA data is too sketchy once you get back a little ways, and even in recent years we see different data with pretty significant differences.

 

Unfortunately, it's difficult to construct a trend line just for reliable land surface data. We can look at the long term trends for places that have seen little in the way of land changes, but those only reflect a small portion of the earth.

[StudentOfClimatology]

You mean like this?

Depends on the dataset you choose to use. GISS and UAH don't have as much of a slowdown in warming. On the other hand, RSS, HADCRUT4, have a much more noticeable one, while NCDC is in the middle of the pack.

While we don't know with certainty which solution is correct, the differences are in fact greater than the supposed margin of error within each dataset. So, someone is wrong, and my hunch is the truth lies somewhere in the middle.

[nflwxman]

Depends on the dataset you choose to use. GISS and UAH don't have as much of a slowdown in warming. On the other hand, RSS, HADCRUT4, have a much more noticeable one, while NCDC is in the middle of the pack.

While we don't know with certainty which solution is correct, the differences are in fact greater than the supposed margin of error within each dataset. So, someone is wrong, and my hunch is the truth lies somewhere in the middle.

Recent evidence suggests arctic bias means the true short term trend is much closer if not slightly higher than GISS since 1998.  Cowtan and Way (2013) shows this very clearly.  In my mind, it makes more sense to infill the data with the UAH data set than ignore it entirely as Hadley does.  

 

Regardless- this is all within the margin of error, which is why short term trends are often meaningless.

[StudentOfClimatology]

Recent evidence suggests arctic bias means the true short term trend is much closer if not slightly higher than GISS since 1998. Cowtan and Way (2013) shows this very clearly. In my mind, it makes more sense to infill the data with the UAH data set than ignore it entirely as Hadley does.

Regardless- this is all within the margin of error, which is why short term trends are often meaningless.

The Arctic isn't the only problem, though. GISS uses ERSST3 for ocean skin temperatures, while HADCRUT/NCDC use HADSST/Reynolds. This is actually responsible for most of the divergence between GISS and HADCRUT/NCDC. I verified this by supplementing satellite data over the Arctic domain into all three datasets, and found the arctic data actually made very little difference post-2005, if any.

GISS was actually closer to NCDC/HADCRUT before it made the switch to ERSST. I don't know why GISS did this, as I personally think it's a less efficient incorporation method.

[ORH_wxman]

The Arctic isn't the only problem, though. GISS uses ERSST3 for ocean skin temperatures, while HADCRUT/NCDC use HADSST/Reynolds. This is actually responsible for most of the divergence between GISS and HADCRUT/NCDC. I verified this by supplementing satellite data over the Arctic domain into all three datasets, and found the arctic data actually made very little difference post-2005, if any.

GISS was actually closer to NCDC/HADCRUT before it made the switch to ERSST. I don't know why GISS did this, as I personally think it's a less efficient incorporation method.

 

 

The Cowtan and Way method too has its own issues...they either kringe land data or infill satellite data over polar ocean areas rather than use OISST data. I'm not sure why they would choose to kringe land temps or use data from 3,000 meters high over water where satellite SST is available....while this didn't make a huge difference in the arctic (which is where most of the commentary was focused), it makes a significant difference over the southern ocean...it severely lessens the southern ocean cooling trend when you replace with TLT data or kringing land data.

 

Basically, all methods have their own idiosyncrasies that can cause divergence in difference regions that may or may not affect the overall global temp calculation.

 

 

As mentioned before, the different methods can be debated until we are all blue in the face, but the underlying differences are generally less than the margin of error. That is why trying to differentiate one year at 0.65 and another at 0.62 is not really productive or all that meaningful.

[chrisf97212]

The HADSST3 data shows the two warmings fairly well. I'm curious about what may have been the culprit behind the 1910-1945 warming:

 

What thoughts do you have on the deeper ocean's impact, if any, on the top 100 meters? 

 

http://ocean.mit.edu/~cwunsch/papersonline/heatcontentchange_26dec2013_ph.pdf