While there’s news of ocean heat content in the Atlantic being pumped up by “leakage” from the Indian Ocean, and NOAA proclaims that La Niña is back, Bob Tisdale finds that the global ocean heat content trend since the turn of the 21st century is flat. Worse than that, it widely diverges from climate models predicting a continued rise in OHC.
2nd Quarter 2011 NODC Global OHC Anomalies
by Bob Tisdale
The NODC updated its Ocean Heat Content Anomaly data to include the 2nd quarter 2011 data. (And they also updated their Thermosteric Sea Level Anomaly data, which is not discussed in this post) I will provide a more detailed discussion as soon as the KNMI Climate Explorer is updated with the 2ndquarter 2011 Ocean Heat Content data, which should be later this month.
THE GRAPHS
Figure 1 is a time-series graph of the NODC Global Ocean Heat Content Anomalies from the start of the dataset (1st Quarter of 1955) to present (2nd Quarter of 2011). The quarterly data for the world oceans is available through the NODC in spreadsheet (.csv ) form (Right Click and Save As: Global OHC Data). While there was a significant increase in Global Ocean Heat Content over the term of the data, Global Ocean Heat Content has flattened in recent years.
Figure 1
And as many are aware, Climate Model Projections of Ocean Heat Content anomalies did not anticipate this flattening. Figure 2 compares the ARGO-era (2003 to present) NODC Global Ocean Heat Content anomalies to the GISS Model-E Projection of 0.7*10^22 Joules per year. The linear trend of the observations is approximately 7% of the trend projected by the model mean of the GISS Model-E.
Figure 2
The source of the 0.7*10^22 Joules per year GISS Model-E ensemble-mean trend was illustrated, clarified, and questioned in the post GISS OHC Model Trends: One Question Answered, Another Uncovered.
HOW MANY MORE YEARS UNTIL GISS MODEL-E CAN BE FOUND TO HAVE FAILED AS A PREDICTOR OF THE IMPACTS OF ANTHROPOGENIC GREENHOUSE GASES ON OCEAN HEAT CONTENT?
I asked the above question in Figure 2. It’s a rewording of the question asked by Roger Pielke Sr., in his post 2011 Update Of The Comparison Of Upper Ocean Heat Content Changes With The GISS Model Predictions. There he notes:
Joules resulting from a positive radiative imbalance must continue to be accumulated in order for global warming to occur. In the last 7 1/2 years there has been an absence of this heating. An important research question is how many more years of this lack of agreement with the GISS model (and other model) predictions must occur before there is wide recognition that the IPCC models have failed as skillful predictions of the effect of the radiative forcing of anthropogenic inputs of greenhouse gases and aerosols.
As far as I’m concerned, they have already failed for numerous reasons. I have illustrated and discussed in past posts how:
1. ENSO is responsible for much of the rise in Ocean Heat Content for many of the ocean basins,
And as far as I know, these are natural contributors to the rise that are overlooked by the GISS Model-E. This was further illustrated and discussed in Why Are OHC Observations (0-700m) Diverging From GISS Projections?
NOTES ABOUT THE ARGO-ERA GRAPH
There will be those who will attempt to dismiss the divergence between model projection and observations shown in Figure 2. Tamino tried to downplay the divergence in his post Favorite Denier Tricks, or How to Hide the Incline. I responded to Tamino with my post On Tamino’s Post “Favorite Denier Tricks Or How To Hide The Incline”. And there may be those who believe 2004 is a more appropriate year to use as the start of the ARGO-era OHC data, so for them, I illustrated how little difference it makes whether the ARGO-era starts in 2003 or 2004 in the post ARGO-Era Start Year: 2003 vs 2004. Note that there are two GISS Model-E projections illustrated in the sole graph in the post ARGO-Era Start Year: 2003 vs 2004. The one at 0.98*10^22 Joules per year, identified as Hansen/Pielke Sr., was found to be in error. This was discussed in the post GISS OHC Model Trends: One Question Answered, Another Uncovered.And of course, there is the fact that natural variables, which are not accounted for by the GISS Model-E, are major contributors to rise in Ocean Heat Content, as discussed in the four posts linked in the previous section.
DATASET INTRODUCTION
The NODC OHC dataset is based on the Levitus et al (2009) paper “Global ocean heat content(1955-2008) in light of recent instrumentation problems”, Geophysical Research Letters. Refer to Manuscript. It was revised in 2010 as noted in the October 18, 2010 post Update And Changes To NODC Ocean Heat Content Data. As described in the NODC’s explanation of ocean heat content (OHC) data changes, the changes result from “data additions and data quality control,” from a switch in base climatology, and from revised Expendable Bathythermograph (XBT) bias calculations.
Kurt in Switzerland: In the following, I’ve copied and pasted the answers I provided to the same questions at my website
Kurt in Switzerland says: “How long have there been sufficient sensors to come up with a semi-accurate calculation of an “ocean heat content” and is there rough agreement on the values thus derived?”
The deployment of ARGO floats ramped up strongly in 2003 and the oceans were well covered early in 2004. To answer the second part of your question, the ARGO data is corrected by researchers to account for sensor problems and the like. The corrections are not always the same.
You asked, “Is there peer-reviewed literature on this, also comparing ocean temperature and climate?”
There are so many, all I can suggest is that you search Google scholar:
http://scholar.google.com/
You asked, “Which scientists are at the cutting edge of this subject?”
This would help with your searches. Scientists whose papers are often referenced in other scientific studies are Willis, Levitus, and Ishii and Kimoto.
You asked, “Also: is there some agreement between scientists studying the ocean heat content and its reflection on / ability to “predict” climate change?”
I have found no evidence that climate model hindcasts/projections represent reality. Refer to:
http://bobtisdale.wordpress.com/2011/04/10/part-1-%e2%80%93-satellite-era-sea-surface-temperature-versus-ipcc-hindcastprojections/
And:
http://bobtisdale.wordpress.com/2011/04/19/492/
Kasuha says:
Honestly, the progress in 1984-1994 is flat in pretty much the same way as in the last five years.
I don’t see the global heat content evolution being flat yet. It slowed down considerably and reduced its variability, yes. But it’s not flat yet.
Yes, it will start to warm up again, some time. Perhaps soon.
But you have to remember that the warmist case is not a consistent trend of the last century. There position is that warming is accelerating, and in an alarming fashion. A return to a small warming will not do them any good.
It will take quite a leap to get up to the those model predictions. I think we can assume they are busted.
David Smith says: “Bob, does table T1 in Levitus show a temperature increase from 1955 to 2005 of 0.17C? That equates to the 15×10*22 J increase in heat content, if I understand the table. “
Which Levitus paper are you referring to? The numbers you’re showing do not correspond to the numbers in Table T1 in Levitus et al (2009) linked in the post.
You wrote, “Surely I am misreading that.”
Referring T1 in Levitus et al (2009), which showed a rise in OHC from 1955 to 2008 of 15.9*10^22 Joules equating to a temperature rise for the top 700 meters of 0.168 deg C, I don’t believe you’re misreading that. That’s my interpretation of that table, and my interpretation is based on the assumption that T1 in Levitus et al (2009) is similar to Table T1 in Levitus et al (2004), where the units and bases are identified:
ftp://ftp.nodc.noaa.gov/pub/data.nodc/woa/PUBLICATIONS/grlheat05.pdf
You wrote, “I’m not suggesting that their work is biased, just that they have ca-ca data with which to work. Do they offer error bars anywhere?”
The only place I know of error data being provided by the NODC is on the data attached to their “Basin time series” webpage:
http://www.nodc.noaa.gov/OC5/3M_HEAT_CONTENT/basin_data.html
The third, fifth and seventh columns are Standard Errors.
Fascinating stuff Bob [as usual], I find imho, browsing figures for OHC though of great academic interest, is not what we should be investigating more thoroughly.
The changes in the PDO with reference to the dynamics which drive the El Nino and La Nina cycles must be understood – because these babies and their dramatic switching and consequential climatic temperature differential drive our weather nuts.
Strewth! The Australians would appreciate a heads up on a severe La Nina but then, so would the rest of the world.
Lastly, I have to agree with Kurt, if we were able to accurately and easily quantify the OHC, then this would enable a better understanding and measurement of world temperature fluctuations, we are a long way from that though.
Ammonite says: “Starting from 2003, are you able to clarify what proportion of the individual model runs (models ER and EH) that compose the ensemble mean fall above and below the 2nd Quarter 2011 endpoint?”
No. The GISS Model-EH and Model-ER data (hindcasts and projections) for OHC are not available in a easy-to-use format
You asked, “Also, what error bars are applicable to the Argo sampled data?”
I have not run across an answer to your question. You’d have to research that through the ARGO website, since the NODC OHC data continues to use other sampling methods during the ARGO era, as far as I know,
Wonderful post, as usual, Mr. Tisdale. Thanks much.
Ocean Heat Content ‘measurements’ pre Argo aren’t worth much – buckets over the sides of ships, etc.
If the Earth’s climate is warming, heat must be accumulating in the oceans, and it clearly isn’t. At least within reach of Argo.
A classic case of observation (measurement) disproving theory.
To answer the question, suppose we had started looking in 1977. Then it isn’t until the late 90s that we see a definite uptrend in the data. Yet we know it was warming during that time. So the answer for how long until things are falsified, is at least another 10 years, and even that doesn’t falsify. We just have some missing heat that needs to be found.
Philip Bradley says: “Ocean Heat Content ‘measurements’ pre Argo aren’t worth much – buckets over the sides of ships, etc.”
Expendable Bathythermographs (XBTs) are a little more practical than attempting to measure ocean temperature at depths of 700 meters with a bucket. The XBTs had their problems but they been in use for a good number of decades before ARGO.
Athelstan. says: “The changes in the PDO with reference to the dynamics which drive the El Nino and La Nina cycles must be understood…”
The PDO is an aftereffect of ENSO.
http://helios.izmiran.rssi.ru/cosray/months.htm
Funny how if you drew trendlines tracing the troughs of the cosmic ray flux, it would correspond nicely with the upward trend, then flattening of the ocean heat content.
Correlation, causation… I know, I know, but still…
Surely this is not a flattening at all. Rather it is a clear indication that the pre-Argo record is not valid. The probability that the real flattening started some years before 2003 argo has to be high 90s. What’is wrong with you scientists? If I had happily handed in that graph as showing anything significant about ocean heat my prof would have sent me down to home economics. But that was pre- post normall science. I guess we can splice the dow jones industrials to he ocean heat data and have peer reviewed discussions about it.
It would be nice to have Figure 3 of Leuliette, E.W., and J.K. Willis. 2011. Balancing the sea level budget. Oceanography 24(2):122–129,
doi:10.5670/oceanog.2011.32.
This shows the sea level change over the same period using Jason’s 1 and 2
http://www.tos.org/oceanography/archive/24-2_leuliette.pdf
I swear there is a lag in Argos temp vs sea height.
Thanks Bob,
I was aware that buckets haven’t been used since the 1940s.
It’s a pet issue of mine that a large proportion of the widely referenced temperature change since 1880 (GISS) and CRU equivalent are based on the unreliable method of sticking a thermometer in a bucket of water hauled up the side of a ship.
Were this more widely known, there would be rather more scepticism about claims of warming over the last century.
Meanwhile, in the main stream land of insanity … SECONNNNNND HOTTESSSSST SUMMMMMERRRRR EVERRRRRRR!!!
2. A change in sea level pressure is likely the cause of the upward shift in North Pacific Ocean Heat content during the late 1980s,
Is that not something that is portrayed as being ‘in tandem’ with a change in the Antarctic? In other words, when the pressure rises over Antarctica, the pressure falls over the N. Hemisphere, and vice-versa. All of which would make the upward shift in N.Pacific OHC late 80’s a natural phenomenon. Solar-induced changes would be on top of, and in addition to, this natural phenomenon, whether positive or negative.
Oh, this post is crap, a vastly oversimplified view of a really complicated subject.
Interesting discussion in the comments at the link below. Which despite its name is a pro-AGW site.
It gets into a issue that interests me,
If OHC isn’t increasing, why are sea levels still rising? The 2 should move in tandem to the extent sea level rises are caused by thermal expansion.
The theory being floated by the pro-AGW camp is that (recent) sea level rises are caused by ocean mass increases, resulting primarily from ice melt.
Thus preserving one of the core AGW predictions, that global warming will cause dangerous sea level rises.
A bit too ad hoc for my taste. More likely IMO is serious problems in either OHC measurements or sea level rise measurements. My money being on the latter.
http://www.skepticalscience.com/A-broader-view-of-sea-level-rise.html
My specific point is, that whenever two estimates of measure location are to be compared to one another, even in the absence of a physical model for their underlying meaning, it is critical to have available the variance for these estimates. I see a lot of discussion here of same or excessive or less or more, but no identification of variance. Can’t tell if two things are different unless and until variance is assessed. Where is it? Moreover, these models are HIGHLY non-linear, as is to be expected from physical phenomena which are essentially fluids.
I would recommend a careful reading of the basics at http://www.realclimate.org, and a respect gains from the details of oceanographic matters, sketched at http://www.whoi.edu.
@Philip Bradley,
“Sea level rise” is not a uniform thing. It is complicated by matters like gravitational attraction of oceanic waters, Corriolis effects, glacial rebound, and so on. It is implausible, if not impossible, with this degree of hetereogeneity in physical causes, to discern effects using information available in measurements alone. Those measurements — and especially their variability — needs to be grounded in physical models.
Moreover, specific impacts of sea-level rise, as in other phenomenological manifestations, such as ice melt, or extreme weather, are only to be experienced at the extremes, not at typical behaviors. That is a statistical phenomenon, almost independent of the physics, where extreme quantiles of a distribution exhibit and are more sensitive to population-wide forces more than the median tends to be.
Bob Tisdale says:
September 8, 2011 at 2:59 pm
Kurt in Switzerland says: “How long have there been sufficient sensors to come up with a semi-accurate calculation of an “ocean heat content” and is there rough agreement on the values thus derived?”
The deployment of ARGO floats ramped up strongly in 2003 and the oceans were well covered early in 2004. To answer the second part of your question, the ARGO data is corrected by researchers to account for sensor problems and the like. The corrections are not always the same.
The expectation of increase in OHC due to theoretical radiative imbalance at TOA is so strong, that the ‘adjustments’ conceal a bias forced by that expectation.
As Peter Berenyi demonstrated in this post
http://tallbloke.wordpress.com/2010/12/20/working-out-where-the-energy-goes-part-2-peter-berenyi/
The radiative balance probably went negative several years ago. That post also demonstrates a likelihood of a bad splice in the data between XBT and ARGO in 2003. The jump in the data looks unphysical to me, given the obvious solar influence which has become more apparent since then.
Josh Willis’ initial findings from ARGO probably did have too much of a fall, but it has likely been overcompensated for. I strongly suspect OHC has been falling slightly since 2002, not rising slightly. This is also evidenced by falling SST as measured by AMSRE. We are just over the peak of the curve, it’s downhill from here.
The only way increased back radiation (has it been measured?) can heat the ocean is via air temperature, not by direct absorption. This is a very slow way to affect the heat content of a four mile deep body of water.
Too slow to have been the cause of the increase in OHC in the ’90s. That would be produced by reduced low cloud over the tropics, a real empirically measured quantity, rather than a theoretically modeled hypothesis.
http://www.climate4you.com/images/HadCRUT3%20and%20TropicalCloudCoverISCCP.gif
Jan Galkowski says: “Oh, this post is crap, a vastly oversimplified view of a really complicated subject”
Did you read the post or only look at the pictures, Jan? Try starting with the opening paragraph of my post. And since you prefer RealClimate’s discussions on the same subject matter, then you’ll have to wait for December or January, which is when they typically offer their model-data comparisons such as:
http://www.realclimate.org/index.php/archives/2009/12/updates-to-model-data-comparisons/
Isn’t that odd, Jan? The Ocean Heat Content model-data comparison portion of that post at RealClimate is “a vastly oversimplified view of a really complicated subject” as well, yet I don’t find a comment from you on that thread advising Gavin his post is crap.
Have a good day.
Thank you for your reply Bob Tisdale. Without assessing either the variability of individual ensemble runs or the error bars on OHC estimation it seems very difficult to determine if the discrepancy between the model ensemble mean and measured OHC is significant. Others have pointed out prior periods of “flat” OHC behaviour extending for over 8 years. My suspicion is that an 8 or 9 year timeframe is simply too short to draw valid conclusions with respect to model accuracy.
Jan Galkowski says:
September 8, 2011 at 9:28 pm
@Philip Bradley,
“Sea level rise” is not a uniform thing. It is complicated by matters like gravitational attraction of oceanic waters, Corriolis effects, glacial rebound, and so on. It is implausible, if not impossible, with this degree of hetereogeneity in physical causes, to discern effects using information available in measurements alone. Those measurements — and especially their variability — needs to be grounded in physical models.”
Wow that sounds the sort of system that one can accurately measure globally rising at 3.2mm a year and have the confidence to adjust by .03mm a year to account for isostatics ——not.
Philip Bradley
I don’t know if you ever caight my article on sea surface temperatures?
http://judithcurry.com/2011/06/27/unknown-and-uncertain-sea-surface-temperatures/
You are right of course, SSTs collected by bucket are nonsensical and any attempts to quantify global SSTs back to 1850 or so is pointless and are a major flaw in the IPCC reports.
As someone above pointed out we had very little computing power even in the 1980’s and in consequence- whilst Bob may have his own ideas as to when we start to have a proper idea as to what SSTs are doing- I suspect that anything prior to the mid 1980’s should be challenged and discarded other than in very localised places and over very limited time scales. .
Nice article as always Bob.
Tonyb