ON Tuesday, I suggested at this blog that I was not convinced by a story from Josh Willis, a scientist at NASA’s Jet Propulsion Laboratory, explaining why and how he corrected data showing global ocean cooling. The title of my blog post suggested that Dr Willis had changed the data to fit output from computer models. Dr Willis has responded, via Fred Singer, explaining that the correction was made, not on the basis of computer output, but on the basis of high quality temperature and pressure measurement from ocean buoys.
“The XBT data were not changed to fit the models. They were changed to agree [with] other higher quality temperature observations from CTD casts, a select group of Argo floats with high quality temperature and pressure measurements, and to be consistent [with] satellite-based observations of the changing height of the ocean surface. The improved agreement with climate models was as much of a surprise to us as anyone else.
In fact, the motivation to look for such problems in the XBT data did not come from comparisons with models, but rather from a comparison with CTD data by Gouretski and Koltermann in their 2007 GRL paper. Gouretski and Koltermann noted that the bias in the XBT data changed over time. After looking into the bias ourselves, we determined that the cause of the bias was a gradual drift in the fall rate of these instruments. By treating the bias as a problem with the fall rate and comparing the XBT data with CTD, Argo and satellite altimeter observations, we were able to estimate the bias and remove it.
Once this data error was corrected, estimates of ocean warming over the past 40 years were much smoother, and the large “bump” in the 1970s and 80s more or less disappeared from the record. This was something of a relief because the “bump” in the heat content record was not only absent from climate models, there was no evidence for it in any other records and no good explanation could be found for its cause. The reason for that is obvious now: because it wasn’t real.”
Thanks to Dr Willis for the clarification and my apologies for any confusion. I have changed the title of Tuesday’s blog post.
Jennifer Marohasy BSc PhD is a critical thinker with expertise in the scientific method.
Surely the lesson is to give institutional scientists some consideration for integrity and professionalism. The quick draw response to suspect sleight of hand as the norm is far from rational.
On the subject of integrity and professionalism, we often hear about sceptics are not to be trusted if they have been on the pay-roll of oil or mining companies.
Does it work the other way as well?
http://margosmaid.blogspot.com/2008/11/on-money-trail-with-tony-jones.html
I think Jennifer is to be congratulated for fessing-up so quickly and comprehensively. Well done. It gives me a lot of faith in the other information that I gather from time to time on this site. Wish that everyone in the debate was trying to be so transparent.
Barf. Did you come down in the last shower Graham?
While not being a scientist, I do not believe that this precludes rational thought and analysis. On reading the explanation above it still appears to me to be a conscious attempt to make specific and limited data match the assumed hypothesis.
Pikey
It’s nice of Dr. Willis to respond to your post. When I read his earlier explanation I got the same impression as what he gave you in the correction. Although I felt like he leaned a little more heavily on getting agreement with the sea level data at that time. But I would now be curious to know this. Since according to the University of Colorado sea level data there has been no sea level rise for the past three years. Will the recent ocean heat content data agree with this, or will the sea level data be adjusted upward to match the newly adjusted heat content data.
Jennifer:
Am I correct in assuming that the data that Dr. Willis made his modifications to were for a publication that went only as far as 2004?
If so, is there a record of ocean heat content since 2004, and can we get at it?
How does that record line up with sea level readings since 2004?
Observations that back other observations…. Now that is science.
Now as to whether they are significant for the AGW hypothesis is another thing entirely.
But a big hat tip to Dr. Willis. He explains his methodology where the Hanson’s and Mann’s of the world will not…. Perhaps, can not.
…. See Debate does work. Jennifer hits a home run for the good guys of science.
Okay, I did a little checking. Apparently the first Willis et al paper came out in 2004 and included data up to mid 2003. This report had a chart showing an increase in ocean heat content. Another came out addressing what was at first thought to be cooling between 2003 and 2006. The errors that were discovered showed that the cooling was not there. However, they showed no warming either.
This is the abstract from the Willis, Lyman, Johnson, Gilson paper.
“Two significant instrument biases have been identified in the in situ profile data used to estimate globally integrated upper-ocean heat content. A large cold bias was discovered in a small fraction of Argo floats along with a smaller but more prevalent warm bias in eXpendable BathyThermograph (XBT) data. These biases appear to have caused the bulk of the upper-ocean cooling signal reported by Lyman et al. (2006) between 2003 and 2005. These systematic data errors are significantly larger than sampling errors in recent years, and are the dominant sources of error in recent estimates of globally integrated upper-ocean heat content variability. The bias in the XBT data is found to be consistent with errors in the fall-rate equations, suggesting a physical explanation for that bias. With biased profiles discarded, no significant warming or cooling is observed in upper-ocean heat content between 2003 and 2006.”
Note the last sentence.
I would actually agree with Willis that the other data over that period (2003-2006), as much of it as I know, would be consistent with an ocean that was not warming or cooling. I do believe, however, that the last two years (using both surface temp and sea level data) should show a small amount of cooling. If this is the case, then we would have no increase in the ocean heat content from 2003 to 2008. And that is not consistent with the models. But it is consistent with the surface temp. From the discussion I have seen between Willis and Pielke on this subject, I get the impression that Willis thinks the deep oceans (beyond the sensors) are still warming. I think that the sea level measurements for the past 3 years would contradict this.
So the oceans are warming?
SJT,
The oceans have been warming – at least up until the last few years. Read the explaination above from Tilo as to what might be happening now.
You tosser SJT didnt you read the last sentence on the previous post which quotes Wilis at al, namely
“With biased profiles discarded, no significant warming or cooling is observed in upper-ocean heat content between 2003 and 2006.”
I don’t think any model projects no increase in the upper ocean heat content for more than 4 years – so if the non-warming continues we can say observations have departed form model projections, and indeed they may already have.
The models can’t predict periods as short as four years, and have never claimed to.
SJT you’re not reading what I said. Model GFD2 CM2.1 supplied by Willis runs from the early 1980s to about 2011 has short periods of non-warming, 4 years at most. The model suggests monotonic warming from 2005 to 2008, following a respite from about 2002 to 2005, a short respite to 2009, followed by more monotonic warming.
If the non-warming for 2003 to 2006 continues into 2008, and beyond, then that is already 5 years. The Barnett et al model (Science, 2001) didn’t seem to show any years of zero heating.
As Willis points out, this doesn’t erase 50 years of warming, but one is left wondering if and how long the non-warming will continue.
Very nice explanation for the XBT data adjustment.
Did he forget his statement about looking for the BIAS causing the last tiny bit of cooling left after this?
Has he expended half the effort in searching for possible WARM bias?? Oh yeah, we KNOW it is warming so no reason to check on that!!
SJT,
according to the CORRECTED data the oceans have NOT been warming since 2003. This is STILL in disagreement with both sea level rise measurements (stopped rising about 1.5 years ago) and the models. THIS explains why he is still looking for cooling bias. He is yet to support the models and/or match sea rise data. This is where the poorly founded sea rise adjustment gets interesting!!!
The recent model runs initialised with recent data, that show a temporary halt in warming, are somewhat matched by this data (of course the initialisation INCLUDES this data!!) As usual, after the break, the model temps start back up. Will the ocean heat content be fully adjusted to match by then??
Finally, what happened to the warming in the pipeline?? Wasn’t the pipeline the ocean stored excess heat?? I get a kick out of the idea that there is somehow a STOREHOUSE of excess heat below 2000 meters, where the ocean temps would appear to be approaching 0c, that can affect the climate for 10’s of years!!
http://www.21stcenturysciencetech.com/articles/ocean.html
http://www.worldclimatereport.com/archive/previous_issues/vol6/v6n16/feature1.htm
SJT:
“The models can’t predict periods as short as four years, and have never claimed to.”
This is true. But considering the fact that sea level has fallen for the past three years that period may well be six years now. We have sea level data up to March of 08 at this point in time. Six years is probably still too short a period for the models. But outside of resolving the issue of model correctness, I think that we can say with some confidence that there has been no ocean heat content increase for almost 6 years. We can also say that if this continues, the models will be in jeopardy. Warmers like Gavin Schmidt claim that we don’t yet have the total effect of the CO2 increase that we have put in the atmosphere. This argument is based upon an ocean heat sync effect. But if the ocean is acting like a heat sync for increasing CO2 forcing, then it’s temperature should at least be increasing, even if the increase is lagging. The fact that it is not increasing sheds some doubt the “lagging effects” theory.
One more thing. If we want to say that the 11 year flat surface temperature trend is there because the heat is being absorbed by the deep ocean, then we have a problem, because heat being absorbed at any level of the ocean should show up as a rise in sea level. And it’s not.
SJT says:
“The oceans have been warming – at least up until the last few years. Read the explaination above from Tilo as to what might be happening now.”
“What might be happening”???!
So, we should ignore the indications of the empirical data and accept any suggestion of “what might be happening” even when Tilo (who reported the suggestion) concludes:
“I get the impression that Willis thinks the deep oceans (beyond the sensors) are still warming. I think that the sea level measurements for the past 3 years would contradict this.”
You really do have to admire true faith.
Richard
Yes, but, the height of the oceans is still questionable. NASA says satellites are only accurate to +/- 3cm loking at a 3mm change per year. And the tide guage sea levels are only rising at about 1.8mm per year. The claimed acceleration in sea level rise appears to be due to compairing guages in the past to satellites now (see graph in NASA article http://earthobservatory.nasa.gov/Features/OceanCooling/page5.php
So making the data fit sea level will not ensure accuracy. And if XBT data is not accurate why not throw it out like he did with Argos data that he did not like? Just use the data from bottles that he likes better? I am still missng what the data from Argos alone shows over the past 5 years. Were any Argos data thrown out because it was too warm or not variable enough? There are just too many assumtions made here for me to be confident in the findings.
>Yes, but, the height of the oceans is still questionable. NASA says satellites are only accurate to +/- 3cm looking at a 3mm change per year.
Gary the central limit theorem means that when you have thousands of observations with an accuracy of 3cm the accuracy of the average is far far greater than 3cm.
All those trained in statistics will be able to verify this for you.
You can be quite confident that NASA knows a thing or two about the accuracy of its satellites and reasonable methods for interpreting the data.
David, The accuracy of the NASA satellite estimate of GLOBAL SEA LEVEL is +/- 3.3 cm according to NASA http://sealevel.jpl.nasa.gov/mission/jason-1.html Needless to say that is alot of water spread over the entire surface of the oceans. I don’t know how much warming it would take to expand the oceans to raise 3.3 cm. But since we only have half of expansion caused by thermal changes, and it is only 3mm rise per year it must be more around a half of a degree. So I will have to continue to be concerned about the level of assumptions been made.
The individual observations are accurate to 3cm. The average of individual observations is far more accurate again.
By all means, be sceptical, but the very most basic of mathematics shows this is a false reason to be sceptical of the sea level data.
David we are all familiar with the saying ” lies damn lies and statistics”
I know many AGW believers are scathing of John Daly, but the following link is well worth a look in terms of showing the huge complications involved in measuring sea level change from satellite.
http://www.john-daly.com/altimetry/topex.htms.
the conclusion is…
“As an example of the best technology available in space monitoring, TOPEX-Poseidon has proved to be a boon to ocean research, not so much because of its sea level averaging, but because of numerous other uses to which this technology has been applied, from detection of El Niño to the tracking of ocean currents and space monitoring of wind speeds through sea state analysis.
But while T/P has proved successful in these other areas, its use as a global sea level monitor is highly restrained by the limitations on its accuracy once millimetric scales are sought. To suggest that changes in the global level of the heaving oceans can be determined to a resolution of tenths of a millimetre, or even a centimetre, is to make unsustainable claims about the inherent capabilities of the technology itself. Statistics cannot compensate for the limitations imposed by the electronics and by radiation physics and claims to the contrary are simply not credible.”
It is well worth a read, he points out “believers” are happy to discard MSU satellite data that does not show the “warm spot” but are happy to believe the TOPEX/ POSEIDON data.
problems include- height of the satellite, footprint size, sea state, atmosphere errors, wavelength of pulses, ocean area coverage.
Richard:
“So, we should ignore the indications of the empirical data ”
The emperical data for the last five or six years shows no warming. The “what might have happened” applies to the confusion between Willis et al 2004 and Willis et al 2006. Stop grandstanding and just read what I said about it – then bring on your objections.
From MIT http://ocean.mit.edu/~cwunsch/papersonline/sea_level_5may2007.pdf At best, the determination and attribution of global mean sea level change lies at the very edge of knowledge and technology. The most urgent job would appear to be the accurate determination of the smallest temperature and salinity changes that can be determined with statistical significance, given the realities of both the observation base and modeling approximations. Both systematic and random errors are of concern, the former particularly, because of the changes in technology and sampling methods over the many decades, the latter from the very great spatial and temporal variability. It remains possible that the database is insufficient to compute mean sea level trends with the accuracy necessary to discuss the impact of global warming – as disappointing as this conclusion may be. The priority has to be to make such calculations possible in the future
“Surely the lesson is to give institutional scientists some consideration for integrity and professionalism. The quick draw response to suspect sleight of hand as the norm is far from rational.”
Don’t be an idiot. No such conclusion follows. Not only are you wrongly extrapolating the specific to the general…. But it appears that the journalist who wrote the story got the wrong impression. I wouldn’t be too quick to let Willis off the hook. I’d want to see what the journalist has to say about it. Whether he is willing to take the blame for this impression he gave. And the explanation doesn’t look full-proof anyway.
If you read the nasa page where Willis’ work was discussed, you would see that the problem was to reconcile a number of apparently contradictory measurement. Which is what they did. Consistency is important.
–Hobgoblin
The reason given for the XBT data problem makes no sense at all……”we determined that the cause of the bias was a gradual drift in the fall rate of these instruments.” These are instruments, presumable all of identical design, weight, cross section etc pulled behind vessels at a defined speed (presumably) and importantly they are disposable so they are not reused to cause issues with and degredation of performance on subsequent use. How can these suffer a bias ‘over time’ due to a ‘gradual drift in fall rate’?…..someone has to explain that one too me. As there are many variables involved; exact speed of vessel, ocean currents, water density etc; their accuracy must be within certain limits anyway. This sounds like the type of reason you come up with when your clutching at straws!!