Listen to this PodcastMore 60-Second Earth
Her Majesty's Ship Challenger set sail in 1872. Stripped of her guns and outfitted for science, her mission was to sail around the globe sampling as she went.
Among other scientific triumphs, the Challenger gathered the first global set of ocean temperature readings, more than 260 in all. The British expedition measured from the surface to a depth beyond 900 meters.
In 2004 a set of drifting buoys began to make similar measurements. There are now more than 3,000 of these floats bobbing in the world's seas, collecting oceanographic information.
Comparing the data sets, separated by more than a century in time, reveals that, yes, the ocean is warming. On average, the global ocean is warmer by roughly 0.6 degrees Celsius at the surface and 0.1 degrees at depth. The analysis appears in the journal Nature Climate Change.
The extra heat trapped by the pollution from more than a century's worth of coal and other fossil fuel burning is beginning to reach the briny deep. That will have impacts from the survival of sea life to global rates of rainfall. Climate change dead ahead.
—David Biello
[The above text is a transcript of this podcast.]
See what we're tweeting about
40 Comments
Add Comment@Tm lsn,
Reply | Report Abuse | Link to thisWhen can we expect your paper? I ask because you must know that simply expressing incredulity about a finding is not a rational critique. Then of course that would require you to actually read the paper.
Link to the Peer Reviewed Paper:
Reply | Report Abuse | Link to this135 years of Global Ocean Warming Between the Challenger Expedition and the Argo Programme:
http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate1461.html
If you go you to the site you can see that they provide error bars for their findings.
.
Thanks for the links.
Reply | Report Abuse | Link to thisI think 'tmb lsn' was asking about the error margins for the sounding temperature data collected by the Challenger.
I don't have $18 to spend or any other way to access the Nature Climate Change article, but weren't the error bars you referred to derived for the new temperature data, not the original Challenger data?
I was able to freely access the supplemental data consisting of the temperature sounding data collected by the Challenger: http://www.nature.com/nclimate/journal/vaop/ncurrent/extref/nclimate1461-s1.pdf
BTW, I notice that much of the sounding temperature data was not collected (missing data). I couldn't assess the impact of that missing data since I couldn't access the complete report. I also can't determine if there was any effort to assess the different temperature measurement methods employed or to normalize the two sets of data...
@JtDwyer,
Reply | Report Abuse | Link to thisIf you read the abstract they give error bounds for the measurements at various depths. I think we can both acknowledge that tmb isn was not simply displaying ignorance but a open hostility and gross ignorance of climate science.
Btw, if you E-mail the authors of a paper and politely ask for copy, a lot of the time they will oblige. Another way is to go to a university or big city library access the database and download the copy.
@ Jtdwyer,
Reply | Report Abuse | Link to thisForgot to add that you can read about the history of sea surface temperatures measurements by one of the relevant experts here:
Revisiting Historical Ocean Surface Temperatures:
http://www.realclimate.org/index.php/archives/2011/07/revisiting-historical-ocean-surface-temperatures/
Of Buckets and Blogs:
http://www.realclimate.org/index.php/archives/2008/06/of-buckets-and-blogs/
It is an interesting story of adjusting for all sort of confounds.
Yes, I try to ignore acrimonious intent - it's bad for my health.
Reply | Report Abuse | Link to thisI was somewhat curious - thanks for pointing out that the error margins were addressed in summary in the abstract. However, error margins were only given there for the temperature differences (warming) between the two datasets, inferring to me (since none were exhibited in the Challenger source data) that they were the product of confidence intervals for the new data only.
The abstract also states: "Systematic errors in the Challenger data mean that these temperature changes are a lower bound on the actual values," again inferring to me that no normalization was attempted on the basis of any specific knowledge of the "systematic errors in the Challenger data."
The report should be very interesting, since the deep fluid oceans can absorbs so much more energy that land can. It does seem oceans and land would be generally exposed to identical amounts of solar radiation on the basis of their surface area, but land surfaces would quickly reradiate large amounts of heat, while the oceans would likely act as a thermal ballast, slowly reradiating heat in the form of evaporation and current flow (more likely affecting long term regional weather and climate).
It's also very interesting that "Warming in the Atlantic Ocean is stronger than in the Pacific." It would be even more instructive if co2 levels in the atmosphere above the Atlantic were greater than those above the Pacific...
At any rate, I appreciate your suggestions, especially that some author might supply a copy of a research report (with proper permissions, of course) - for some that's never occurred to me! Thanks again.
Jt Dwyer Says: It's also very interesting that "Warming in the Atlantic Ocean is stronger than in the Pacific." It would be even more instructive if co2 levels in the atmosphere above the Atlantic were greater than those above the Pacific...
Reply | Report Abuse | Link to thisTrent Says: I am afraid the explanation for that is a bit more prosaic than that. The Pacific is on average about a 1,000 meters more deep than the Atlantic.
Ocean:
http://www.eoearth.org/article/Ocean
Luddite
Reply | Report Abuse | Link to this1. Title of the article is very misleading, if not an outright lie. They are not comparing 135 years of data...rather two sets of data 135 years apart. BIG difference.
Reply | Report Abuse | Link to this2. Any change in the data from 1872 to present is completely meaningless UNLESS you can compare it with temperature data PRIOR to 1872...Have any?
Jt Dwyer Says: Thanks.
Reply | Report Abuse | Link to thisBTW, more meaningful than average depth in determining their temperature difference is the total volume of water they each contain: ~324 million cubic kilometers for the Atlantic; ~622 million cubic km for the Pacific. Their volumes should generally indicate how much heat they could absorb while their surface areas would better indicate how much radiant heat they were exposed to: ~105 million sq. km for the Atlantic; ~165 million sq. km for the Pacific.
Reply | Report Abuse | Link to thisOf course, other factors would also be influential, such as a much greater percentage of the Pacific's volume is located in the tropics while the Atlantic narrows there, between Africa and South America, and the total surface area of land that drains into the Atlantic is about 4 times that of the Pacific.
(source: respective wikipedia entries)
@Rottncf: . Any change in the data from 1872 to present is completely meaningless UNLESS you can compare it with temperature data PRIOR to 1872...Have any?
Reply | Report Abuse | Link to thisTrent: You just made a unsupported assertion. Care to support exactly why you must have data prior to 1872 to tell if their has been a change in temperature since 1872? What does that even mean? Do you think at all about what you say?
@JtDywer Says: . Their volumes should generally indicate how much heat they could absorb while their surface areas would better indicate how much radiant heat they were exposed to.
Reply | Report Abuse | Link to thisTrent Says: I am sorry to get pedantic here but the above is not necessarily true. You could have a body of water that had greater volume but had much less depth thus exposing a greater portion of it to irradiance. That of course is not the case here of course.
Whilst this is interesting, I do feel it is a bit of a stretch to assume that man is responsible for the heating of the deep ocean based only on two sets of data separated by 135 years. The accuracy of the original data sets cannot be verified, and a difference of only 0.1°C at depths below 900m (~3000ft) could be easily explained by errors in the original instruments and recording process.
Reply | Report Abuse | Link to thisFurther, why do they not consider other impacts on the temperature where the differences are recorded, such as cyclic changes in deep ocean currents or even the depth of the deep ocean floor as a result of tectonic activity.
To blame every little difference on the activity of man, regardless of the fact we need to clean our act up, is myopic, unscientific and bordering of fraud. I appreciate that the intent of the Nature Article is not intended to mislead, but the conclusions are as far fetched as the sky demon myths of the Palaeolithic era that we now call God!!
It is about time we had real honesty put back into environmental science, less hockey sticks and less jumping to conclusions by either side of the fence so that the real picture can be gleaned, but that would not suit Political and Science funding criteria so I shall not hold my breath.
Personally, I think that humanity's increasing industrial activity and enormous (7-fold) population growth over the past ~200 years are significantly affecting environmental and climatic conditions, likely including ocean temperatures.
Reply | Report Abuse | Link to thisThat being said, I also think that every scientific study should be questioned not just by professional peers with similar interests but external groups who may have equally valid but differing perspectives.
You've expressed very well several potential issues (also alluded to by other commentators) with this study. As you say, the report is interesting but the evidence presented is hardly compelling by scientific standards.
By the way, as I understand this was not a peer reviewed article published in Nature but a letter published in the more parochial journal nature climate change (stylishly, no caps in the title).
The preceding comment was intended to be a response to JMFranklin's comment, above.
Reply | Report Abuse | Link to thisMoreover ...
Reply | Report Abuse | Link to thissame time of year?
similar weather conditions?
any change in prevailing currents in 135 years?
any assessment of accuracy of the instruments used in 1872?
It's like going into Manchester on a rainy weekend in December, returning on a dry day in July and declaring that the climate has changed. Bad science par excellence!
The earth is how many BILLIONS of years old? And we're going to base a trend on 135 years of information? Don't think that sample will result in a high degree of confidence....
Reply | Report Abuse | Link to thisThis is very interesting regardless of the veracity of the conclusions drawn. I have tremendous admiration for the scientists and explorers of that era. Also people like Cook, Bering & the early polar explorers.
Reply | Report Abuse | Link to thisThere is still a great wealth of information that will eventually be mined from all sorts of sources. One source of data that I often think about is the records kept by beekeepers. There are many long-term records that show seasonal variations and variations in the flowering times of plants of interest to beekeepers. It is a pity when researchers succumb to the natural human instinct of finding patterns in the evidence they are examining that conforms to their preconceptions. It is crucial that when evidence like this is examined that it truly is peer review. It is so easy with minor inclusions or omissions in the margins are so small to completely skew the results.
An experiment that I would like to see that to my knowledge has never been carried out would be to have the luminosity of the moon, Earthshine, recorded and correlated with the relevant positions of the oceans. It may be possible to detect variations in the radiation also resulting from cloud cover using this method.
JMFranklin Says:hilst this is interesting, I do feel it is a bit of a stretch to assume that man is responsible for the heating of the deep ocean based only on two sets of data separated by 135 years.
Reply | Report Abuse | Link to thisTrent: Did you read the article? This paper is not about the attribution of climate change, but is an observation of it.
JMFranklin:It is about time we had real honesty put back into environmental science, less hockey sticks...
Trent Says: The presumption here is that something nefarious has been going on with climate science. I am afraid you have fallen for anti-science propaganda. Are you aware that the Hockey Stick has been confirmed again and again by a dozen different studies using different proxies and methodologies. For you to repeat such falsehoods tells me that you need to get a different source of science news.
@Mlbbbchbill,
Reply | Report Abuse | Link to thisI do wish you would think about what you say. Most people who are adults recognize that the same phenomena can different causes. This is like saying that since forest fires happened way before humans ever appeared that arson is impossible.
Typical liberal Trent, you cannot argue the point so you belittle the person or start the name calling...the reason there are more acres of forrest today in the United States than ever before is our ability to put out fires. That has nothing to do with the statistical relavence of 135 out of 4,000,000,000.00 which is right next to nothing.
Reply | Report Abuse | Link to thisoh no now the chicken little how something new to run around and yell about. fact is neature will always be chaning that is her nature, pun intended. but if it were getting colder then they would change their tune and be screeming about that.
Reply | Report Abuse | Link to thisThe data (at least as presented in this article) can't be used to draw any meaningful conclusions regarding ocean temperature fluctuation in the 135 year period since the Challenger voyage or any time before it. As such, it is an interesting set of isolated data from a set of instruments that cannot be calibrated against modern thermometers. The mean differences in temperature can not be assessed as both statistically significant and scientifically meaningful, and should be viewed as anecdotal evidence in support of the general thesis that greenhouse gasses are the primary cause of the warming trend we are (or at least have been) experiencing. That evidence does not preclude other, perhaps more significant factors that may also contribute to long-term or intermediate term climate change.
Reply | Report Abuse | Link to this@Mlbbchbill,
Reply | Report Abuse | Link to thisCan you explain how the following is name calling.How about flaw in the logic?
"I do wish you would think about what you say. Most people who are adults recognize that the same phenomena can different causes. This is like saying that since forest fires happened way before humans ever appeared that arson is impossible."
Going to answer or do intend to keep sidetracking the conversation?
Interesting points. I have read the abstract and then donwloaded the additional information, which is the data sets that were used to draw the conclusions indicated in this article. The main glaring scientific error that I see so far is the 'Argo temperature data (°C, 2004-2010) interpolated14 to the time of year and location of the Challenger stations.'
Reply | Report Abuse | Link to thisSo there are a couple of issues (that I am going to probably have to spend $18 to figure out if they were addressed):
1. The data from the Challenger Stations probably have some associated measurement error. Was that error quantified and if so how does it relate to the difference in the two data sets.
2. The Argo data has been interpolated and seasonally corrected to the Challenger data points which opens that data set up to potentially huge amounts of error. My questions regarding that are: (1) what was the interpolation methods used (and were multiple methods applied) and at what grid scale and (2) what was the amount of seasonal correction that was applied and what base line was that correction determined from?
I am going to go ahead and purchase and read the article but I think that these conclusions are not going to bear up under scrutinty.
Regarding your issue 1. - that is mentioned in the abstract. As I mentioned earlier, the abstract states: "Systematic errors in the Challenger data mean that these temperature changes are a lower bound on the actual values."
Reply | Report Abuse | Link to thisThis infers to me that no normalization was attempted on the basis of any specific knowledge of the "systematic errors in the Challenger data." The researchers seem to indicate that the 'actual' temperature differences between the two datasets is likely greater than reported...
Trent1492, do you really think you make yourself look good, or add to the argument with a personal attack, that is not really mature nor is it helpful.
Reply | Report Abuse | Link to thisYour ask if I read the article, the answer is Yes I did, and Yes I do understand the science. I do not think I need another source of Science Data, unless your suggesting that such lofty journals as Nature, The Monthly Notices of the Royal Astronomical Society, Astrophysical Journal, American Institute of Aeronautics and Astronautics, Atmospheric Chemistry and Physics, Geophysical Research Letters, Journal of Applied Meteorology and Climatology, Physical Review Letters and Advanced Composite Materials are not credible sources because as a subscriber I receive them at every publication...and holding a BSc in Astronomy, with a PHd in Planetary Atmospheres I actually understand the majority of what is published, added to this I hold an MSc in Electrical Engineering and currently design, construct, install and maintain the cellular network Infrastructure here in the UK...so I understand the engineering behind the science equipment too.
Trent, if you read my comments correctly, you would see that I freely acknowledge that man needs to clean his act up, we do not use our homes as toilets, and we should not do the same with our environment either, the halting of pouring poisons into the atmosphere is clearly important and this should not be minimised, but the claims made by the Anthropogenic Climate change lobby are bordering on an obsession that is actually harming the whole argument. Every time something is found to be different from what it was before the conclusion is always that man is responsible, and this is a ridiculous situation.
That man can clearly impact the environment is not in doubt, but the impact we are having on the global climate is, the case is not proven as many claim, and in fact there is as much scientific and peer reviewed evidence against the claims than there is in it's favour.
The claims that the majority of scientists agree with this is incorrect, the fact is the majority simply refuse to condemn because those that have often suffer professionally and even suffer abuse in the normal press. The intimidation needs to stop and an open and honest discussion needs to be had in which all are safe and free to speak openly. Simply attacking and vilifying individuals who do not agree with your view point is akin to the Church condemning the likes of Tycho Brahe, Copernicus and Galileo.
@JmFranklin,
Reply | Report Abuse | Link to thisYou have offered no defense of your comment. You claimed to have read the article but I having trouble finding where attribution for the warming is found. Care to cite the relevant passage?
All you have done is offer a word salad and no coherent message. And you have made several false and contradictory claims. For example here you say,"..and in fact there is as much scientific and peer reviewed evidence against the claims than there is in it's favour."
Then you say in the next paragraph, "...the claims that the majority of scientists agree with this is incorrect, the fact is the majority simply refuse to condemn because those that have often suffer professionally and even suffer abuse in the normal press."
So which is it? Are the scientist being repressed or are they free to publish just as much they wish? You can not have it both ways. And no your claims are in factual error:
Here is a listing of every climate science paper published since the 19th century up to 2011.
http://www.skepticalscience.com/Interactive-History-of-Climate-Science.html
Notice that by 2011 there have been 186 skeptical papers published and 2409 pro-agw published. Go to the site and you will find for each year the list of publications.
Your second claim that the scientific community is evenly split is again factually incorrect:
Examining the Scientific Consensus
on Climate Change:
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CDEQFjAA&url=http%3A%2F%2Ftigger.uic.edu%2F~pdoran%2F012009_Doran_final.pdf&ei=mA96T6qtAsKYiAKg6cRb&usg=AFQjCNFW7gAEg5cN1_dx1xWE7Cpqu81x4Q&sig2=uaAfjYlYnvjF30BJDOAxjA
The paper listed above is a survey of the scientific community and found that %82 of the scientific community agreed with the statement that the Earth is warming and that humans are causing it. As the scientist relevant expertise increases so does the the agreement till it reaches %96.7
As for this supposed harassment of skeptical scientist? In your mind only. It is the climate scientist who are being harassed by law suits, fake scandals, and death threats. Just Michael Mann or Kathraine Hayhoe.
I cannot agree that because there are systematic errors that the reported differences are only the lower bound. If there is systematic error in the Challenger data set and what is being reported is only likely to be the lower bound then what is the likely upper bound?
Reply | Report Abuse | Link to thisIf the systematic error is biased is some manner from a Gaussian distribution then maybe the statement in the abstract is correct. However, as I said before I need to purchase the article and read it because I have a feeling that thermometer technology 135 years ago probably had a bit more error than 0.6 degC. It is the distribution of the error in the original data that will show where the computed difference lies between the two data sets.
In any case even if the Challenger data points are 100% accurate and the Argo data points, where they were measured, are 100% accurate there is the issue of the interpolation and seasonal correction. While I applaud the author's idea of attempting to spatially match the two data sets I believe that the computed differences will be swamped by the potential error in the spatial and seasonal correction of the two data sets.
Since the two data sets are freely available and I have Excel I am going to see what lies in the statistical distribution of the differences.
In the end the real problem here is that we do not have enough data over a long enough time to draw any solid conclusions as to whether or not the oceans are indeed warmer now than they were in the past. It is the same with the atmospheric temperature data sets as well. We have very little reliable data to go on so we are stuck with the incomplete analyses with too little data.
Anyhow, I was unable to figure out how to purchase the paper today so I will try again tomorrow!
I agree with your assessment plan, although I don't see where any confidence intervals can be obtained for the original Challenger data (unless their measurement equipment & procedures can be tested).
Reply | Report Abuse | Link to thisBTW, I think the letter can be purchased at:
http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate1461.html
by clicking on the 'Buy Now' button, or at:
https://secure.nature.com/store/svc/article/cart?action=add&articleId=nclimate1461
You may have to register for a free id first, but I also found the process very confusing (just poking around). For customer service, try (General Site Visitor):
Email: registration@natureny.com
Tel: 1-866-363-7860
Sorry - the preceding comment intended to be a response to evosburgh.
Reply | Report Abuse | Link to thisThanks for the info on purchasing. I also found the process confusing and I plan to contact customer service.
Reply | Report Abuse | Link to thisI have loaded the data to both Excel and my preferred time series analysis program and done some preliminary statistical analysis. In performing that analysis I have found some interesting things which are based upon the assumption that the data is 100% correct:
Considering all of the data points I see that the min and max are -8.3 and 9.1 with a mean of 0.26 and the std deviation is 1.1. In this the 50th percentile is 0.25 and the 10th and 90th are -0.45 and 1.20.
In looking at the displays of the data (which I wish I could show here) is reasonably well constrained within a 1.65 degree range and has Gaussian distribution.
It is pretty clear that the shallower the measurement the warmer the water appears. I am almost willing to totally neglect the 0m reading as it is going to be too affected by the local daily solar input to provide a meaningful comparison. When you take the average of the mean change for the 100 to 1000m readings you find that the change is around 0.26 degC over the past 135 years.
Up to this point I have assumed that the data is 100% accurate. However, the run of the mill modern digital thermometer has an accuracy of +/-0.2 degC when new. So let's for arguments sake assume that the historic and modern data sets both suffer from this range of error in their readings (which is a stretch because there is no way that thermometers from 135 years ago were as accurate). Then we have a range of +/-0.4 degC, which is probably one standard deviation meaning that there is a 64% chance that the readings are within that range, and we apply that to our analysis.
If we apply that error to the existing average between 100 and 1000m we find that the apparent warming of 0.26 degC is within the first standard deviation of the measurements (+/-0.2 degC for each data set) that this says to me is that unless the difference is more than 0.4 degC then the result lies within the resolution of the measurement.
In order to make a reading that makes statistical sense we would have to use modern thermometers and then take station logs at fixed locations in over a gien period of time, many years, in order to make a reasonably sound judgement about the warming, or cooling, of the world's oceans.
To the extent that I understand your assessment, I certainly agree that the computed temperature differential is not likely significant. Well done!
Reply | Report Abuse | Link to thisCritically, we're also assuming that the logged depths of the Challenger temperature readings were accurate...
You do bring up an interesting point, which I neglected: the potential depth error and ensuing temperature reading error in the data sets.
Reply | Report Abuse | Link to thisWe also assume that the Argo depths (which are determined by pressure) are also correct. I am somewhat less concerned about the depth measurement because the gradient in the Challenger data set is 0.015 degC/meter from 100 to 1000 meters (0.016 for the Argo set). Therefore they would have to be off by something like 17 meters to induce an error equivalent the reported temperature differential between the data sets.
I am going to go out on a limb and say that there depth measurement error is probably no more than a half of a meter at 1000m if we apply a modern depth measurement accuracy (around 0.05% * depth). You do bring up an interesting point that I forgot which is the potential depth error and ensuing temperature reading error in the data sets.
So ultimately even if the Challenger depth measurements are off by twice what we would expect from modern accuracy the difference at 1000m is around 0.015 degC which is about 1/17th of the reported differential which is still way inside the accuracy of the thermometers.
You're probably right, but I don't think I've seen how depth was determined for the Challenger readings. I can imagine the simplest methods (cable length, etc.) might be off by significant amounts due to deep water currents & such.
Reply | Report Abuse | Link to thisFirstly Challenger didn’t work in metric they were British and as such worked in nautical terms, so they used feet, and Fathoms.
Reply | Report Abuse | Link to thisThe means by which they measured depth in those days was by a length of rope or Cable, which was 720 feet long, with a lead weight with knots at 6 foot marks, or Fathoms, I believe in the case of the Challenger they used a specially made cable with coloured markings every Fathom usually yellow, with a red one every tenth marker.
So unless they were measured carefully when originally made they could be out an inch every Fathom which after 1,000 meters could add up to at least a couple of feet.
Another factor to consider is the time of day that each of the measurements were taken as well, were the corresponding modern data taken not only at the exact spot as the Challenger but at the same time of day.
But regardless of the two comparisons it really is pointless comparing two sets of data 135 years apart without corresponding data in between to give a gradient.
Thanks for the info & insights. For direct comparison the readings should be taken as you say at the same spot at the same time of day and (surely it's been mentioned) on the same date.
Reply | Report Abuse | Link to thisMost critical is your final point - well done! It is the variation among many observations that helps to identify the likely contributing factors that collude to produce those variations.
Here is climate scientist Roger Pielke, Sr.'s comment on this article:
Reply | Report Abuse | Link to thishttp://pielkeclimatesci.wordpress.com/2012/04/03/comments-on-the-scientific-american-article-135-years-of-records-reveals-deep-ocean-warming/
He wasn't complimentary:
"Perhaps the article is an April Fools joke as it was published on April 1st. Unfortunately, it is not and represents another example of science being misused to advocate a particular viewpoint on the climate issue."
Thanks for the link. Pielke's comments seem quite reasonable to me. However, I am surprised that a climate scientist apparently didn't obtain the nature climate change article, directing his specific remark to this SA report.
Reply | Report Abuse | Link to thisPielke's only specific comment: "To equate ocean temperature anomalies from data obtained from one ship in 1872 with the large Argo array is absurd." He does provide a link to a map of the ARGO monitor network: http://www.argo.ucsd.edu/About_Argo.html