Comments on the Nature Article “Climate Change: The Case of the Missing Heat”

UPDATE: I finished writing this post and published it at my blog Climate Observations about the same time that Don Easterbrook’s post Cause of ‘the pause’ in global warming was published at WattsUpWithThat.

This post includes an overview of the mistakes people make when they attempt to use the Pacific Decadal Oscillation data as a metric for the dominant mode of variability in the Pacific Ocean. So it contradicts Easterbrook’s post. More detailed discussions of what the PDO data represents and what it doesn’t represent can be found in the posts here, here and here. If you for some reason would like to argue about it on this thread, please see my comments on the Easterbrook thread starting here.

# # #

The Nature article Climate Change: The Case of the Missing Heat by Jeff Tollefson has received a lot of attention at WattsUpWithThat and around the blogosphere. Much of Tollefson’s article was a simplified explanation of the findings of Trenberth and Fasullo (2013) An apparent hiatus in global warming? We’ve discussed that paper already. See the posts Trenberth and Fasullo Try to Keep the Fantasy Alive and More on Trenberth and Fasullo (2013) “An Apparent Hiatus in Global Warming?”

Anthony Watts was right to call attention to Tollefson’s article; it presented an argument and illustration that had been presented years ago by skeptics, an argument and illustration that had been dismissed by global warming enthusiasts. And Tollefson is linking it to Trenberth and Fasullo.

One point needs to be made right from the get go. As far as I could tell, no one interviewed for the Nature article claimed the hiatus in surface warming was not happening. That article is an attempt to explain the growing differences between the model simulations and measured global average surface temperatures. And the more I look at it, Tollefson appears a bit skeptical at times.

One problem with Tollefson’s article: it only discusses the hiatus in surface temperature warming. It does not address the absence of warming of the Pacific Ocean and North Atlantic to depths of 2000 meters, for the period of 2003 to 2012. See Figure 1. Only the South Atlantic and Indian Oceans show warming to depth. Figure 1 is from the post here.


Figure 1

Manmade greenhouse gases cannot select which ocean basins they decide to warm.

And the Tollefson article, unfortunately, offers as much misinformation as information.

The first problem, as noted above, was the article does not address the “missing heat” in the depths of the Pacific over the last decade, and that is where the heat would have to be appearing if Trenberth and Fasullo (2013) were correct.

The second problem has to do with using Pacific Decadal Oscillation (PDO) data as a metric for the strength, frequency and duration of El Niño and La Niña events. It’s the wrong dataset.

El Niño and La Niña events are the primary focus of the article. Tollefson acknowledges this with his statement:

And here, the spotlight falls on the equatorial Pacific.

But then he presents Pacific Decadal Oscillation (PDO) data as the metric for El Niños and La Niñas. He’s likely taking his cue from Trenberth and Fasullo (2013) for that.

El Niño and La Niña events take place in the tropical Pacific, but the Pacific Decadal Oscillation dataset is an abstract form of sea surface temperature data derived from the extratropical North Pacific, not the tropical Pacific. See Figure 2.

02 PDO Region

Figure 2

The PDO data basically represent how closely the spatial pattern of the sea surface temperature anomalies of the North Pacific (north of 20N) resembles the pattern created by ENSO events (where the spatial patterns are warmer in the east and cooler in the central and western North Pacific during El Niños, and cooler in the east and warmer in the central and western North Pacific during La Niñas). Because that spatial pattern of sea surface temperatures in the North Pacific is also impacted by sea level pressures (and their interrelated wind patterns), the Pacific Decadal Oscillation data are not a true representation of the frequency, magnitude and duration of El Niño and La Niña events.

For more information about what the PDO represents, and more importantly what it doesn’t represent, see the posts here, here and here.

03 The Pacific s Global Reach

Figure 3

The big surprise in the Nature article was the illustration comparing the Pacific Decadal Oscillation (PDO) data with the global surface temperatures titled “The Pacific’s Global Reach”. I’ve included it as my Figure 3. It compares the JISAO PDO data to global surface temperature anomalies. The text in the illustration clearly states:

During periods when the PDO index is positive and the eastern Pacific is warm, global temperatures have risen quickly. During spells when the PDO index is negative, the warming has stagnated.

Additionally, with respect to the illustration, the article reads:

An analysis of historical data buttressed these conclusions, showing that the cool phase of the PDO coincided with a few decades of cooler temperatures after the Second World War (see ‘The Pacific’s global reach’), and that the warm phase lined up with the sharp spike seen in global temperatures between 1976 and 1998 (ref. 4).

This, of course, strongly suggests that the warm phase of the PDO is responsible for part of the warming during the period from 1976 to 1998. If the PDO can stop the warming, then the PDO had to have contributed to the warming. That had been a skeptical argument for years. I can recall seeing similar graphs more than 5 years ago. In fact, it was so popular SkepticalScience wrote a post about it here back in 2008.

Unfortunately, there are no mechanisms through which the PDO (as defined by the JISAO dataset) can cause global surface temperatures to vary. The sea surface temperatures of the North Pacific north of 20N are inversely related to the PDO data.

Those mechanisms exist in the tropical Pacific and portray themselves as El Niños and La Niñas. Sea surface temperature-based indices for the strength, frequency and duration of El Niños and La Niñas do exist. In fact, Trenberth and Stepaniak (2001) Indices of El Niño Evolution (online edition here) recommended using the sea surface temperature anomalies of the NINO3.4 region of the equatorial Pacific (identified as N3.4 in the paper). See the illustration here for the location of the NINO3.4 region. Trenberth and Stepaniak (2001) used the HADISST dataset, with the base years of 1950-1979 for anomalies. Unfortunately, Trenberth and Stepaniak never really provided a scientific justification for using the base years of 1950 to 1979. So we would be better off using the entire term of the data (1923 to 2013) as the base years for the NINO3.4 sea surface temperature anomalies.

The NINO3.4 sea surface temperature data are realistic index for strength, frequency and duration of El Niño and La Niña events. So let’s look at the average NINO3.4 sea surface temperature anomalies for the two warming periods and the two hiatus periods. See Figure 4. We’ll use the periods defined by Trenberth and Fasullo (2013):

The monthly time series (Figure 8) readily reveals the multidecadal regimes of the PDO (given by the black line) with positive phases from 1923 to 1942 and 1976 to 1998, and negative phases from 1943 to 1976 and after 1999.

Curiously, the periods presented by Trenberth and Fasullo (2013) do not agree with those presented by Tollefson in his “The Pacific’s Global Reach” illustration. That aside, my Figure 4 shows that during the two warming periods of 1923 to 1942 and 1976 to 1998, the average sea surface temperature anomalies for the NINO3.4 region were well above zero. This means that El Niño events dominated that period. Global surface temperatures should warm during those periods because the tropical Pacific is releasing more heat than “normal” from the tropical Pacific and redistributing more warm water than “normal” to adjacent ocean basins. And during the two hiatus periods of 1943 to 1976 and 1999 to present, the average NINO3.4 sea surface temperature anomalies are below zero. That of course means the tropical Pacific is releasing less heat than “normal” and redistributing less warm water than “normal” to adjacent ocean basins. And during those periods global surface temperatures flattened or cooled slightly.

04 NINO3.4 SSTa w-Period Averages 2

Figure 4

You may be asking yourself, why did Bob go through that exercise when it basically showed the same results as Tollefson’s PDO graph?

The NINO3.4 data are the correct dataset to use and the PDO is not. The PDO does not represent the timing, strength and duration of El Niño and La Niña events. It portrays an aftereffect of El Niño and La Niña events, combined with the influence of the sea level pressure of the North Pacific.

But now we have to consider that the NINO3.4 sea surface data only capture the effects of ENSO on a small region of the equatorial Pacific. They do not represent the processes of ENSO or their aftereffects. And in no way can NINO3.4 sea surface temperature anomalies account for the warm water that is left over from strong El Niño events, which is redistributed to adjacent ocean basins in their wakes.


The article states (my boldface):

Just before the hiatus took hold, that region had turned unusually warm during the El Niño of 1997–98, which fuelled extreme weather across the planet, from floods in Chile and Califor­nia to droughts and wildfires in Mexico and Indonesia. But it ended just as quickly as it had begun, and by late 1998 cold waters — a mark of El Niño’s sister effect, La Niña — had returned to the eastern equatorial Pacific with a vengeance. More importantly, the entire eastern Pacific flipped into a cool state that has continued more or less to this day.

Entire eastern Pacific? The PDO data do not represent the “entire eastern Pacific”. If he’s thinking sea surface temperature anomalies, that’s also wrong. There have been a number of El Niño events since 1999. Those El Niño events occurred in 2002/03…and 2004/05…and 2006/07…and 2009/10. As a result, the average sea surface temperature anomalies [of the East Pacific], using NOAA’s standard base years 1971-2000 for their Reynolds OI.v2 data, for the period of January 1999 to December 2013 are +0.06 deg C, and that’s a far cry from “a cool state that has continued more or less to this day.” See Figure 5.

05 East Pacific SSTa 1999-2013

Figure 5


The article reads:

Second, many researchers have found the opposite pattern in simulations with full cli­mate models, which factor in the suite of atmospheric and oceanic interac­tions beyond the equatorial Pacific. These tend to reveal a trend towards more El Niño-like conditions as a result of global warming.

Climate models don’t simulate ENSO properly, so this is irrelevant. See Guilyardi, et al. (2009) “Understanding El Niño in Ocean-Atmosphere General Circulation Models: Progress and Challenges” and Bellenger, et al. (2013): “ENSO Representation in Climate Models: From CMIP3 to CMIP5.” Preprint copy is here.


The following was written for the post Comments on Stefan Rahmstorf’s Post at RealClimate “What ocean heating reveals about global warming”. The discussion applies here as well. I’ve changed the Figure numbers and made a few minor revisions for this post.

In the left-hand graph of Figure 6, we can see that the sea surface temperatures of the East Pacific have warmed little, if at all, over nearly the past 32 years…since the start of the Reynolds OI.v2 sea surface temperature dataset in November 1981. The East Pacific (90S-90N, 180-80W) covers about 33% of the surface of the global oceans. The papers presented by Jeff Tollefson do not address this lack of warming and climate models do not simulate it.

06 E Pac and S Atl-Ind-W Pac

Figure 6

The same hold true for what’s illustrated in the right-hand graph, and they are the sea surface temperature anomalies of the South Atlantic, Indian and West Pacific oceans. The coordinates are listed in the title block. They cover more than 50% of the surface of the global oceans. We can see a number of things in the sea surface temperature data for the South Atlantic, Indian and West Pacific oceans: (1) their long-term warming depends on the El Niño events of 1986/87/88 and 1997/98 (and possibly the El Niño of 2009/10); (2) they do not cool proportionally during the transition to the trailing La Niñas of 1988/89 and 1998-01; but (3) they do cool slightly over the decade-long periods between the El Niño events of 1986/87/88 and 1997/98 and between the 1997/98 and 2009/10 El Niños.

For an introductory discussion of the natural warming of the global oceans refer to the illustrated essay “The Manmade Global Warming Challenge” (42MB).

The sea surface temperatures of the North Atlantic are governed by another mode of natural variability called the Atlantic Multidecadal Oscillation or AMO, which has been overlooked by Jeff Tollefson. The AMO is the reason why the sea surface temperatures there warmed at a much higher rate than the rest of the global oceans from the mid-1970s to the early-2000s. But the sea surface temperature data for the North Atlantic indicate the AMO may have already peaked. See Figure 7. If history repeats itself, and there is no reason to believe it will not, then the sea surface temperatures of the North Atlantic will show no warming and actually cool for a few more decades, assuming the Atlantic Multidecadal Oscillation has, in fact, peaked.

07 AMO May Have Peaked

Figure 7

Note: For further information about the Atlantic Multidecadal Oscillation, see NOAA’s AOML (Atlantic Oceanographic and Meteorological Laboratory) Frequently Asked Questions webpage here, and my blog post here and my introduction to the Atlantic Multidecadal Oscillation here.


The article by Jeff Tollefson suggests the climate science community may be moving in the right direction—that’s a big MAY BE.

But they’ll have to return to basics: ENSO is fueled by sunlight. And for that reference, as I’ve presented numerous times before, we have Trenberth et al. (2002) Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures:

The negative feedback between SST and surface fluxes can be interpreted as showing the importance of the discharge of heat during El Niño events and of the recharge of heat during La Niña events. Relatively clear skies in the central and eastern tropical Pacific allow solar radiation to enter the ocean, apparently offsetting the below normal SSTs, but the heat is carried away by Ekman drift, ocean currents, and adjustments through ocean Rossby and Kelvin waves, and the heat is stored in the western Pacific tropics. This is not simply a rearrangement of the ocean heat, but also a restoration of heat in the ocean.

In other words, the stronger trade winds reduce cloud cover, which, in turn, allows more sunlight to warm the tropical Pacific. So, if the stronger trade winds are causing more warm water to be stacked up in the western tropical Pacific, it’s an increase in sunlight that’s fueling it. This was discussed in much more detail under the heading of DOWNWARD SHORTWAVE RADIATION VERSUS DOWNWARD LONGWAVE RADIATION DURING LA NIÑA EVENTS in the post Open Letter to the Royal Meteorological Society Regarding Dr. Trenberth’s Article “Has Global Warming Stalled?”


newest oldest most voted
Notify of

Again, you are the best! Thanks!


Looks like the AMO index may have peaked and is now declining .You are right. Also I have found that the North Atlantic, North Pacific and the Northern Hemisphere SSTs are all now declining since about 2004/2005.

John Tyler

Actually neither heat nor anything else is missing. When your analytical models are WRONG – as evidenced by their inability to model today’s climate and more damning , the HISTORICAL climate – then an honest researcher must admit that they , at present, do not fully understand what drives the climate.
The only other serious field of endeavor that religiously follows models that are ALWAYS WRONG, is economics. Note that in both climate ” science” and economic astrology ( oops , I meant science) , controlled experiments are impossible and politics predominates. The real world evidence matters not at all.


Your view and analysis has proven to be more correct than the team’s. Keep up the good work.


I’m having a problem with the statement:
“In other words, the stronger trade winds reduce cloud cover, which, in turn, allows more sunlight to warm the tropical Pacific. So, if the stronger trade winds are causing more warm water to be stacked up in the western tropical Pacific, it’s an increase in sunlight that’s fueling it. ”
It seems the trade winds are the cause of two things: 1. More sunlight in the tropical pacific and 2. Water to be stacked up in the western tropical pacific.
The fact that the water being stacked up is warm is due to more sunlight, which in turn is caused by stronger trade winds.
Am I understanding this correctly?

If I were asked to pick a single point on earth that most likely has the greatest effect on global weather and climate, it would be 0 and 90W (Galapagos). This is where El-nino winds, the deep sea Cromwell current, the Panama current, and the Humbolt current meet. These flows are not constant and each has different cycles and those cycles are not constant. Cycles on cycles create extremes in weather and climate. These extremes have an effect globally. I suspect these cycles are also controlling our observed atmospheric concentration of CO2. CO2 is very likely a lagging indicator and not a cause of climate change.

Gail Combs

fhhaynie says: @ January 18, 2014 at 8:24 am….
Do you think some time in the future you could elaborate on that and post it here on WUWT as an article?
For those who are curious like I am this is a map of sea currents from WIKI
The Cromwell Current is also called Pacific Equatorial Undercurrent or the Equatorial Undercurrent and is eastward-flowing and down 300 feet (100 m) under the surface of the Pacific Ocean at the Equator. (again WIKI)
For Galápagos and Panama Current information see link (It has an interesting map and the usual CAGW crud.) The Panama Current is a seasonal warm current.
The Humbolt current “is a cold, low-salinity ocean current that flows north along the west coast of South America from the southern tip of Chile to northern Peru.” WIKI again and a MAP
Since I have an interest in the water going through Drake Passage at the tip of South America, I will add the Cape Horn Current. “…The Cape Horn Current is a cold water current that flows west-to-east around Cape Horn. This current is caused by the intensification of the West Wind Drift as it rounds the cape….” WIKI
My interest is because Antarctic sea ice growth could affect the wind and currents in that area especially at the Drake Passage restriction. This could then affect the cold waters headed up the west coast of South America and that might affect ENSO.
From another source: CHAPTER 31 OCEAN CURRENTS: TYPES AND CAUSES OF CURRENTS (Maritime Safety Information) I start at the begining of the passage which talks of the southern Atlantic. The chapter also has a map that shows that some of the cold water heads up along the west coast of South America.

…That branch of the South Equatorial Current which curves toward the south off the east coast of South America, follows the coast as the warm, highly-saline Brazil Current, which in some respects resembles a weak Gulf Stream. Off Uruguay it encounters the colder, less-salty Falkland or Malvinas Current forming a sharp meandering front in which eddies may form. The two currents curve toward the east to form the broad, slow-moving, South Atlantic Current in the general vicinity of the prevailing westerlies and the front dissipates somewhat. This current flows eastward to a point west of the Cape of Good Hope, where it curves northward to follow the west coast of Africa as the strong Benguela Current, augmented somewhat by part of the Agulhas Current flowing around the southern part of Africa from the Indian Ocean. As it continues northward, the current gradually widens and slows. At a point east of St. Helena Island it curves westward to continue as part of the South Equatorial Current, thus completing the counterclockwise circulation of the South Atlantic. The Benguela Current is also augmented somewhat by the West Wind Drift, a current which flows easterly around Antarctica. As the West Wind Drift flows past Cape Horn, that part in the immediate vicinity of the cape is called the Cape Horn Current. This current rounds the cape and flows in a northerly and northeasterly direction along the coast of South America as the Falkland or Malvinas Current….

We need to know the time series’ of flow rates and temperatures of those currents. Those cold currents contain a lot of CO2 that is emitted as the sun warms them in their travels to and westward along the equator. I’m looking at the reanalysis SSTs and air data but that does not provide sea current flow data.

Hello Bob,
Can you please help me with this question:
IF the Sun is the dominant factor in determining ocean and atmospheric temperature, can one explain the observation that the ~90 year Gleissberg Cycle is in-and-out-of-phase with the ~60 year PDO?
Thank you for all your good work, Allan

Peter Miller

Using alarmist logic, I think I may have found the reason why the South Atlantic and the Indian Ocean are warming up a little, while the Pacific and the North Allantic Oceans are basically flat lining.
Trenberth’s missing heat (these great depths suck this heat out of the rest of the ocean) is hiding in the Puerto Rican Trench, the deepest part of the Atlantic Ocean and in the Marina Trench in the Pacific, the deepest point in the oceans on Earth.
I repeat, I had to use alarmist logic to come to this conclusion.

Very clear Bob, thankyou.

Village Idiot

So much for settled science on this side of the fence


Village Idiot says:
January 18, 2014 at 10:25 am
So much for settled science on this side of the fence.
FB asks: What do you mean?
Science is always sitting ON the fence. That is often a little bit uncomfortable , which is why it is never settled.
Settled science? Slightly oxymoronic ?

Richard M

I have posted the theories of Dr. Willam Gray more than once. His thoughts, as I understand them, are that ENSO variability is driven by the MOC. Hence, we have MOC -> ENSO -> PDO. The MOC also drives the AMO.
The MOC is the giant circulation within the global oceans. It speeds up and slows down due to density differences (and other possible factors) across the entire planet. When the MOC speeds up you get more cold water upwelling and flowing up the west coast of S. America. This cold water enhances the trade winds which reduces the probability of seeing El Nino events. It also means that warm surface water spends less time giving up heat before it is driven to the poles, cools and submerges.
This circulation could also be behind the longer term changes like the MWP and LIA.
As some have indicated the most recent PDO probably ended around 2005-7. Here’s a graphic that ties the phases of the PDO to hadrut4:
As one can see the trends match perfectly. Remove the adjustments made to the data and we are left with a very minor warming trend with variations due to the oceans.
Also note that the warm-up of the Indian and South Atlantic Oceans are also explained. With fewer El Niño events the warm water is pushed by the trade winds into the Indian Ocean which then affects the South Atlantic.

Matthew R Marler

Thank you for a good post. I hope that Mr Tollefson’s article will be the first of many probing particular inadequacies of the consensus views. As always, … .

John F. Hultquist

Gail Combs says:
January 18, 2014 at 9:25 am
fhhaynie says: @ January 18, 2014 at 8:24 am….
“Do you think some time in the future you could elaborate . . .

While we are waiting, here is a good read that speaks to ocean currents and lack thereof; selected for the annotations in the left column:

Gail Combs

John F. Hultquist says…
OH!, an old favorite. Thanks.


This is perfect, love it! I believe it matches perfectly what I believed, but had a couple holes in explaining, for a while!
So just to recap …and correct me if I am wrong – that is really what I am asking for here, after all. Just want to make sure I understand this correctly! (don’t want to be one of those blindly believing a theory I don’t really understand when I get into the weeds, as the other side has done for years)
If I am understanding it correctly, then Ocean Oscillations (which would after-the-fact depict data from the El Niños and La Niñas) and Solar Activity should correlate quite nicely to “Global Warming” (as we have seen in pieces such as
– and I won’t rely on that as 100% Gospel truth, there could be errors, but there can’t be that many I image)
And if the above is correct, and we are currently facing this in reality
– PDO in Cool (meaning overall downward trend) for at least another 20 years
– AMO is likely peaking so becoming cooler (meaning overall downward trend) – a trend which should last roughly 40 years before it goes back into its next Warming trend
– Solar is quite likely going into some form of one of the following
– – – a late 1800s or 70s style lull (at best)
– – – a Dalton type Minimum
– – – a Maunder type Minimum (at worst)
– – – a Sporer Minumum
– – – a Homeric Minimum
Those three things (PDO/AMO/Solar all negative trending) being the case, then a noticeable (and possibly fairly extreme) cooling trend should probably be on our horizon (and rather specifically, most likely to be on the dramatic side in the next 20 years), right?
And if it is, then this
which depicts an overall downward trend for the past 10,000 years is looking like said trend might ultimately be continuing rather perfectly,? A quick eyeballing tells us that trend even says it could/should be a 0.5-1.0 Degree decrease in Greenland, at least, in the coming couple years
…or have I run completely off the reservation at this point?

jai mitchell

why did you pick anomalous periods in figure 4? The handy selection of data points that are abnormally low/high in your ranges indicate a significant selection bias in your processes. I know this is very unscientific, but it seems that your figure 4 periods should correlate to your figure 3 periods if you are trying to say something that is associated with the PDO or Global Warming.


jai mitchell says:
January 18, 2014 at 1:01 pm
We’ll use the periods defined by Trenberth and Fasullo (2013):
The monthly time series (Figure 8) readily reveals the multidecadal regimes of the PDO (given by the black line) with positive phases from 1923 to 1942 and 1976 to 1998, and negative phases from 1943 to 1976 and after 1999.


If we combine the effects of PDO and solar cycles, how much does this explain warming since 1970 ?

Barry Wells

The perception, like the sun, might just be beginning to turn – well worth a view.
Barry Wells

Steve from Rockwood

Thanks Bob. I have a feeling you’re pulling us closer to the truth.


This might sound trivial, but if you dive down 20 feet and stay there in a sea, say to scuba dive, it ain’t half cold without that wet suit. Sometimes you blind me with science, but I do know what I have personally experienced around the world, and sea temps do fluctuate. In Kyrenia in Cyprus, the harbor was much warmer than the sea outside. It was shallower of course. Is that trite. With tides especially Ebb tides, only on a full moon, the seas naturally rise higher onto the land.

Steve Keohane

Thanks for keeping it real Bob.


Has anyone do quantitative analysts. I have and quite honestly you only concentrate on one area, taken on a global scale it is near impossible.