Guest Post by Willis Eschenbach
I took a lot of flak last year for my post saying that the global 50% drop in phytoplankton claimed by Boyce et. al was an illusion. I had said:
So where did the Nature paper go wrong?
The short answer is that I don’t know … but I don’t believe their results. The paper is very detailed, in particular the Supplementary Online Information (SOI). It all seems well thought out and investigated … but I don’t believe their results. They have noted and discussed various sources of error. They have compared the use of Secchi disks as a proxy, and covered most of the ground clearly … and I still don’t believe their results.
In other words, I took my chances on my experience and went way, way out on a limb with my statements. And of course, people didn’t let me forget it.
Figure 1. Life Cycle of Phytoplankton
Now we get these two “Brief Communications Arising“, from Nature magazine (emphasis mine).
Nature Volume: 472, Pages: E6–E7
Brief Communication Arising (April, 2011) Arising from D. G. Boyce, M. R. Lewis & B. Worm Nature 466, 591–596 (2010)
Phytoplankton account for approximately 50% of global primary production, form the trophic base of nearly all marine ecosystems, are fundamental in trophic energy transfer and have key roles in climate regulation, carbon sequestration and oxygen production. Boyce et al.1 compiled a chlorophyll index by combining in situ chlorophyll and Secchi disk depth measurements that spanned a more than 100-year time period and showed a decrease in marine phytoplankton biomass of approximately 1% of the global median per year over the past century. Eight decades of data on phytoplankton biomass collected in the North Atlantic by the Continuous Plankton Recorder (CPR) survey2, however, showan increase in an index of chlorophyll (Phytoplankton Colour Index) in both the Northeast and Northwest Atlantic basins3, 4, 5, 6, 7 (Fig. 1), and other long-term time series, including the Hawaii Ocean Time-series (HOT)8, the Bermuda Atlantic Time Series (BATS)8 and the California Cooperative Oceanic Fisheries Investigations (CalCOFI)9 also indicate increased phytoplankton biomass over the last 20–50 years. These findings, which were not discussed by Boyce et al.1, are not in accordance with their conclusions and illustrate the importance of using consistent observations when estimating long-term trends.
Along with this one:
Nature 472, E5–E6 (14 April 2011)
Brief Communication Arising (April, 2011) Arising from D. G. Boyce, M. R. Lewis & B. Worm Nature 466, 591–596 (2010)
… Closer examination reveals that time-dependent changes in sampling methodology combined with a consistent bias in the relationship between in situ and transparency-derived chlorophyll (Chl) measurements generate a spurious trend in the synthesis of phytoplankton estimates used by Boyce et al.1. Our results indicate that much, if not all, of the century-long decline reported by Boyce et al. is attributable to this temporal sampling bias and not to a global decrease in phytoplankton biomass.
OK, so I was right. The Boyce paper was nonsense, the claimed trend was spurious, plankton biomass is holding somewhere near steady or even increasing, and a number of independent records show that the Boyce et al. paper is garbage built on bad assumptions.
I bring this up for three reasons. The first is to show the continuing shabby quality of peer-review at scientific magazines when the subject is even peripherally related to climate. Nature magazine blew it again, and unfortunately, these days that’s no news at all. It’s just more shonky science from the AGW crowd … and people claim the reason the public doesn’t trust climate scientists is a “communications problem”? It’s not. It’s a garbage science problem, and all the communications theory in the world won’t fix garbage science.
The second reason I posted this is just because I enjoy it when it turns out that I’m right, particularly on a risky statement made with no data and in the face of opposition, and I wanted to enter that fact into the record. Childish, I know, but at least I’m adult enough to admit it.
The third reason is a bit more complex. It is to emphasize the value of actual experience. I didn’t disbelieve Boyce et al. because I had any data. I had no data at all.
What I did have was a lifetime spent on and under the ocean. Phytoplankton form the basis of all life in the ocean. If the phytoplankton had actually gone down by 50%, all life in the ocean would have gone down by 50% … and my experience said no way that was true. Fish catches haven’t gone down like that, numbers of species on the reef and along the coast haven’t gone down like that, I would have noticed, people around the world would have been screaming about it.
So I put my neck on the chopping block, and I trusted my experience … and in the end, despite the people who laughed at me and abused my claims, my experience won out over Boyce’s “science”.
Does this mean that we should always trust our experience over science? Don’t be daft. Science is hugely valuable, and often shows that our experience has misled us completely.
But far too many scientists forget to check the obvious – their own experience. Not one of the Boyce authors thought “Wait a minute … since the oceans live almost entirely off the phytoplankton, if plankton is down by half why haven’t I seen oceanic populations from krill to whales and octopuses dropping by half?” Or perhaps they just didn’t have the experience to check the obvious.
The moral of this story? Well, the moral for me is that trusting my experience over the “science” of high-powered scientists living in an ivory tower far above the ocean worked out well … this time.
But the real moral is that scientists need to pay more attention to the “laugh test”. I know when I first heard the Boyce claim, I busted out laughing … and when our experience is that strong in saying that science is wrong, it’s likely worth checking out.
w.
ADDED LATER For me, there’s a few tests that I apply regularly that seem to not be applied by far too many mainstream AGW supporting scientists. These are the smell test, the laugh test, and the eyeball test.
The laugh test weeds out the worst, like the preposterous claim that plankton had been decreasing by 1% per year for the last century. The “Rule of 70” gives the doubling time for an investment. You divide 70 by the annual interest rate, and that gives the doubling time in years.
The Rule of Seventy gives seventy years for doubling time at 1%, so that means in a hundred years the total biomass of the ocean has decreased by more than 50% … seriously, I laughed. The biomass of the ocean decreased by more than half and nobody noticed until now?
The smell test is more subtle. It depends on the provenance of the information, and the way it has been handled, whether it looks “natural”, the history of the investigators, and the like. While the smell test can’t reject anything, it shows me where to look for something wrong.
The eyeball test is simple. Look at every single dataset. There is absolutely no substitute for the experienced human eye. You say there’s seven thousand of them? Boo hoo. You can’t just make up an algorithm and apply it without seeing what it does to every single station record. If there’s a large number, I just write a program in R that just flashes them on the screen for a second along with an ID number. I just let it roll, and jot down the numbers of the ones that stand out.
Now before anyone starts screaming about computerized checking, yes, they are extremely valuable. I’m a whiz at error-trapping, anomaly finding, and computerized checking in a variety of computer languages.
But computerized checking is only as good as the person who wrote the computerized checks. And until you understand every different way that your data might be contaminated or tainted or erroneous for a host of reasons, you will not be able to write computerized checks to identify those particular errors.
And the only way to do that is to put each and every dataset to the eyeball test. There’s an example of what I mean in my post When Good Proxies Go Bad over at ClimateAudit. (Y’all should definitely visit ClimateAudit, Steve McIntyre is continuing his amazing exposition of the never-ending revelations of the “hide the decline” fandango …)
Smokey,
it would have left 50%, assuming the median is at the mean (i.e. a symmetrical distribution).
At least every other day I encounter a Science Headline that causes me to LOL, as did the Phytoplankton loss story.
I would offer this as another:
http://www.scientificamerican.com/article.cfm?id=ozone-hole-may-have-cause-australian-floods&WT.mc_id=SA_CAT_SP_20110425
Could not you infer, that since Australia has had as large or larger floods in the past, that perhaps the ozone hole has been here for awhile?
More funding for ozone research needed now!
RW
“far too many people who blindly follow peer reviewed articles seem unable, or unwilling, to compare what they are led to believe with what often can be observed in the real physical world around them.”
Apparently it depends on the conclusion the article reaches. If the conclusion is validating of the reader’s own view, they tend to believe it. If it is counter to the reader’s own view, then it isn’t so blindly followed.
I seem to remember an incident a couple of years ago where certain stations in Finland and Russia reported the same temperature data in consecutive months. In other words, one month’s data was reported again as the subsequent month’s data. It turns out that this had happened before but had gone unnoticed. During the middle of summer or winter when temperatures are stable, this might not be noticed at all, it was only noticed in spring when certain stations reported exceptionally cold temperatures because they were reporting March data in April. GISS then noticed and corrected it. They apparently didn’t notice it when the same thing happened in the fall when temperatures were being reported exceptionally high (September temperatures being reported for October) because they “wanted” to see exceptionally warm temperatures and so they assumed those reading were “correct” and they validated their hypothesis.
It was Steve McIntyre who noticed it in the autumn months and the list of “bad” data was expanded on at this site with several more being located.
So the moral of the story is that people tend not to closely check data that validates their expectations.
Even so, if you have data that are suddenly wildly different than data surrounding them, it is always good to give it a close look, one’s own view notwithstanding.
Here is the full exchange in Nature including the Boyce et al. reply.
I’ve put an addendum up at the head post.
w.
Good work Willis but you make one statement I have to disagree with. You say you had no evidence. Not true, you had exceptionally good evidence notably that life in the sea had not gone down by 50%. The chain of inference is extremely short and robust – phytoplankton form the basis for all marine life, if they decreaseso should all marine life. The fact that marine life had not decreased was quite enough of itself to discount the claims.
The instinct of an experienced person is often based on data that should be “obvious” but often is not. Their paper as as bad as the student who is given a problem to calculate the power output of a power station and submits the answer of 50 microwatts without thinking of what the answer would imply.
I remember reading (somewhere lost in time) that excellence comes from applying equal parts of knowledge, skill, and talent. If any of these three get out of balance, excellence suffers. We are a society too willing to presume excellence flows from knowledge alone.
A 1% decrease in phytoplankton biomass per year, every year for 1 century would produce a total drop of 2.7x. By today, 111 years later, phytoplankton biomass would be only 33% of the 1900 biomass. Cetaceans should be lined up outside soup kitchens.
Phytoplankton are also responsible for about half the atmospheric oxygen. I’d guess we’d be seeing a decrease in oxygen content by now, if phytoplankton output was down by ~67%.
Some wisdonm learnt by me in engineering over the years is to identify early on the particulars of what “Governs” for a particular project.
For Willis here I think he knew that plankton “Governs” ocean life.
It comes down to pattern recognition. Animals can do it, computers can’t.
If you want to examine the data, start by graphing it. Too many scientists trust their ‘closed form’ equations using statistics and various kinds of transforms, which don’t present temporal patterns in a way that the MK I Eyeball can analyze.
Look first at the waveform.
Willis: Common-sense and reasoning can only take us so far, usually in the direction of our personal biases. For example, common sense says that the amount of relative humidity of the atmosphere will remain constant as the earth warms from radiative forcing. For science (as opposed to the blogosphere) to progress, we need solid observations and analysis to reject the conclusions from papers like this one.
The peer reviewers did a lousy job with this paper. a) With a gradual changeover from one measurement technique to another, the reviewers should have demanded to see separate trends for each measurement technique. b) What effect does temperature have on the growth of phytoplankton under controlled conditions? If a 1 degC change in temperature causes little change in growth rate, then something besides global warming must be responsible for observational changes. The center of large oceans are unproductive environments because nutrients (particularly iron) are in short supply. Averaging results of a large area means that the large variation of productivity with location is lost.
Well I am going to put myself out on a limb. I reckon ALL the sea and land temp data from GISS, Hadcrut, NDCD, NOAA and the arctic ice data from cryosphere and NCDC has been manipulated to suit the AGW agenda. Basically previous temps have been lowered and recent temps have been artificially increased to show the artificial warming. The ice data has been manipulated by changing the areas they measure in NH and manipulating base averages. Let’s see if I have been right (for future records)
richard telford says:
April 25, 2011 at 1:58 pm
If I recall correctly, you (or someone) made that argument over in the original thread. You are correct, but it is a difference that doesn’t make a difference.
You are correct that the primary productivity is what counts. You could think of it with an analogy from the land. Suppose we are taking a snapshot of the amount of hay in the field. Does that tell us the productivity of that field?
Well, no, because it might be that the field only produces one crop of hay per year, or it might produce three crops of hay in a year … but looking at the snapshot (the instantaneous biomass) can’t tell us which one it is. So you are right, there is a difference between biomass and primary productivity.
The reason that it is a difference that doesn’t make a difference is that the plankton in the ocean isn’t suddenly changing from one crop per year to three crops per year, to use the previous analogy. Plankton doesn’t suddenly get three times as efficient, it does what it has always done.
So as a result, the primary productivity varies roughly proportional to the total biomass. Not exactly, but their claim is that the biomass in 2010 is only a third as large as in 1900 … are you trying to explain that by saying there’s only a third the plankton, but they’re three times as efficient?
Again, the laugh test is in play. You are 100% correct, but that doesn’t explain their erroneous results.
w.
illustrate the importance of using consistent observations when estimating long-term trends
Well Duh…
Willis:
Reminds me of a time when I was a young engineer, arguing about the cause of a problem. An old stationary engineer posited a possible cause for the problem. I expertly, concisely and totally demolished his argument. I made him look like a total fool . . . expect for that small problem of his being right and me being wrong. Science involves debating at times, but the rules are a little different. Just because you win the debate doesn’t mean you are right. Sooner or later, science will out and the right answer will be known.
Well done with the BS meter.
kwik says:
April 25, 2011 at 1:58 pm
I’m not into conspiracy theories (although I was wrong in the case of Climategate), and even that one had no “puppet masters”, just a host of victims of noble cause corruption.
w.
Wisdom is what you have when you have made every possible mistake so many times that you have finally learned no to make them! Alas, some people revere their mistakes and never learn.
Willis,
Would you believe that this planet was a water world greater than 1 billion years ago?
This planet looses .00025mm of water through the atmosphere per year which translates to 2.5mm/10,000 years.
If you decide to follow this math, then a few hundred meters of water has been lost in 1 billion years.
eNihilist says: April 25, 2011 at 1:32 pm “Nice on Dr. Boyce also. Any person who can admit when they’re wrong (listening Team members?) is a stand-up character in my book.”
The brief communications that Willis quoted are not by Dr. Boyce. His reply can be found at http://www.nature.com/nature/journal/v472/n7342/abs/nature09953.html
He maintains that the problems pointed out by the 3 brief communications regarding his paper are minor and do not materially affect the results.
How to confound a scientist:
Ask him
“how do you call an instrument to look at very small things?”
And he’ll answer matter-of-factly “a microscope”
“how do you call an instrument to look at things very far?”
and he’ll answer, obviously, “a telescope”
“how do you call an instrument to look inside objects?”
He’ll think a second and answer “an endoscope” with the tiny smile of someone “in the know”
Then ask him:
“how do you call an instrument that can make you see through walls?”
Then, watch him break a sweat, think of various ray or remote sensing devices… then give him the answer: “a window…”.
Anthony Watts says:
April 25, 2011 at 2:46 pm
When I was a kid, first we plowed, then we disked, then we harrowed, then we planted. I deny categorically that we used Secchi disks.
Life always astonishes me with its tenacity and power. Grass grows up through cracks in the concrete, tree roots split stones, species stubbornly refuse to go extinct …
Thanks, Anthony, and thanks as always for the superb site.
w.
Werner Brozek says:
April 25, 2011 at 2:53 pm
We have little global data on oceanic pH. In addition, pH is an intensive variable, so an “average” is a somewhat tenuous concept.
w.
Smokey says:
April 25, 2011 at 3:03 pm
Around a third.
w.
“It’s a garbage science problem, and all the communications theory in the world won’t fix garbage science.”
That’s true and that’s why they have to roll out Flannery. It’s a brain washing exercise in the same vain as they daily brainwash our children. You can see that they are getting desperate because now in Leftist press and just the other day on a-pac TV the rumblings are indignant that so many people are against the carbon (sic) tax. The elites are lamenting our democracy and think that we should have a government like communist China. That way the govt. could just tell us what to accept! This is about science but it’s now gone way beyond this. We are fighting for our current way of life. Capitalism is purposely being demonized and is under attack here and the USA, socialism is lauded. It’s an agenda which is being played out in most western countries. My grandfather died in WWII fighting such tyranny. He’d be turning in his grave right now.
When reading about the latest high profile publication in climate science, I’m frequently reminded of the term that Phillip Morrison used in his book and old PBS videos on physics: “the ring of truth”. Most of the time the sound is more of a thud than a ring, and I think, Who reviews this stuff?
Good call, Willis. Now about those lizards in Mexico…