Update: see below for the latest bit of continuing obstinate idiocy from Steve McIntyre, plus another diagram, plus another reference.
There are undoubtedly a few knowledgeable people chuckling or groaning over an article Steve McIntyre wrote on his blog last week. In case you are, like me, loathe to visit a blog where the author has a tendency to both conspiracy ideation and alleging the work of scientists is a "scam" and their results "fake", these here are the nuts and bolts (as I understand it).
Steve McIntyre mistook water mass movement for temperature (archived here). Or more correctly, he mistook a measure which was used to indicate water mass movement as being used to indicate water temperature.
If you're wondering how anyone, even a climate science denier, could possibly make such a silly mistake, read on.
Rahmstorf15 and the Atlantic Meridional Overturning Circulation
You may recall the hullabaloo that followed a new paper by Stefan Rahmstorf and colleagues. I've already written about the three protest articles (in almost as many hours) on Anthony Watts denier blog (click here). Now Steve McIntyre is protesting loud and long - and making a huge fool of himself in the process. (He'd already written at least two protest articles that I haven't bothered to read. If they are as bad as this, his third(?), then they'll be very bad indeed.)
Rahmstorf15 was about the Atlantic Meridional Overturning Circulation (AMOC), and how it may have slowed in the late twentieth century (particularly between 1970 and 1990). The authors attributed this slowing to the influx of fresh cold water, primarily from melting of the Greenland ice sheet. The implication is that, because Greenland is going to continue to melt, the AMOC may continue to slow down and cause marked changes in climate up north, and probably in more places, too. The massive ocean currents have a turnover time of around 1,000 years - which is long term as far as humans are concerned. However the impact of any major changes would be felt in the short and medium term (decades to centuries).
To learn more, go read the article by the lead author, Stefan Rahmstorf, at realclimate.org, if you don't subscribe to Nature Climate Change (or don't want to buy the paper itself). I've previously put together a list of other articles on the paper, which are worth reading.)
How Steve McIntyre gets it wrong (again)
I heard about Steve's big blooper from a tip from Bill H in the comments here. So I popped over to Steve's blog (which I don't usually bother with). Steve was thinking that Stefan Rahmstorf and all his colleagues got something upside down. He was comparing a chart in Rahmstorf15, the AMOC paper, with one that he called "Sherwood11". Something about "d15N" data from corals off the coast of Canada.
Steve was in such a rush to show that "scientists don't know nuffin'" that he didn't say what "Sherwood11" was. No link. Nothing. (My first thought was U-NSW's Steve Sherwood, though I hadn't seen him write about corals before.) After a bit of detective work, I discovered Steve's "Sherwood11" was this paper by Owen Sherwood and co. (it's open access).
The Sherwood paper was about a nutrient regime shift in the western North Atlantic (as you can tell from the title: "Nutrient regime shift in the western North Atlantic indicated by compound-specific δ15N of deep-sea gorgonian corals"). The abstract suggests why this research was used by Stefan Rahmstorf in his study. Here's the relevant part from the abstract (my emphasis):
...In the Northwest Atlantic off Nova Scotia, coral δ15N is correlated with increasing presence of subtropical versus subpolar slope waters over the twentieth century. By using the new δ15N-AA approach to control for variable trophic processing, we are able to interpret coral bulk δ15N values as a proxy for nitrate source and, hence, slope water source partitioning. We conclude that the persistence of the warm, nutrient-rich regime since the early 1970s is largely unique in the context of the last approximately 1,800 yr. This evidence suggests that nutrient variability in this region is coordinated with recent changes in global climate and underscores the broad potential of δ15N-AA for paleoceanographic studies of the marine N cycle.
The Sherwood paper was describing changes in nutrient levels over the past 1,800 years or so, as measured by δ15N in a species of coral south of Nova Scotia. From that, the scientists were able to deduce changes in sources of water over time - with more or less water coming from the sub-Arctic compared to the sub-tropics.
BTW, "slope" refers to the continental slope, as shown in this diagram from the US Office of Naval Research:
|Source: Office of Naval Research|
When water from the cold sub-Arctic water dominates the region, it is much lower in nutrients than when warm water from the subtropics dominates. (This is described in Sherwood11 - and there's another description at NOAA, in the context of the North Atlantic Oscillation.)
Below is Figure 1 from Sherwood11, on which I've highlighted the area that the specimens came from. The map also shows the different sources of water, which dominate at different times:
|Figure 1: Map of the study area with location of Northeast Channel. Blue arrows mark the approximate position of the Labrador Current, which transports Labrador slope water (LSW) southward, along-slope. Red arrows show the location of the Gulf Stream. Warm slope water (WSW) occupies the region between the Gulf Stream and continental shelf edge. Temperature and nitrate data were extracted from the 2 × 2 degree grid centered over the Northeast Channel, from 150 to 250 m water depth, where interannual variability is most pronounced (18). Temperature data are from the Bedford Institute of Oceanography hydrographic database. Nutrient data are from the Biochem database of Fisheries and Oceans Canada (44). Source: Sherwood11|
How Steve McIntyre got his correlations wrong
Steve McIntyre mistakenly thought that Rahmstorf was using the δ15N data as a proxy for temperature of the water off Nova Scotia. They weren't. They were using the data (and findings) of Sherwood11 as an indication of the ocean currents. Steve mockingly wrote:
The idea that coldwater corals offshore Nova Scotia can be thermometers for ocean temperature in the subpolar gyre has little more plausibility than the belief that stripbark bristlecones in the distant Sierra Nevadas or contaminated Finnish sediments can be thermometers for the subpolar gyre.
If you're familiar with Rahmstorf15 then you'll be saying to yourself - but the paper was primarily focused on the region of the subpolar gyre, whereas Sherwood11 was about a location just south of Nova Scotia. And you'd be correct. You'll also be wondering why Steve assumed that the δ15N data used by Rahmstorf15 was as a temperature proxy - when both Rahmstorf15 and Sherwood11 were primarily about changes in ocean currents. Sure, Rahmstorf's AMOC index was based on temperature, but it was based on temperature of the sub-polar gyre - not on the temperature of the waters south of Nova Scotia. And Sherwood11 did discuss the temperature of the waters that were more (subtropical) or less (subarctic) rich in nutrients. But temperature wasn't the reason that Rahmstorf15 used the Sherwood11 data.
Steve had his correlations all wrong. If you still haven't caught on why, keep reading.
Movement of water mass, not a temperature proxy
Rahmstorf15 is all about the strength of the AMOC - how it slowed down a lot in recent years, particularly between 1970 and 1990. And it still hasn't got as fast as it was in the first half of the twentieth century. At one stage in the analysis, they used the corals as an illustration of the movement of ocean water mass (not temperature).
But first the diagram - Figure 5, which Steve picked out from Rahmstorf15, which shows the AMOC index (derived by the authors, based on temperature) as well as the annual mean bulk δ15N from Sherwood11 (as an indicator of water mass changes, not temperature changes) - as always, click to enlarge it:
In the text of Rahmstorf15, the authors state (my emphasis):
...The green curve denotes oceanic nitrogen 15 proxy data from corals of the US north-east coast from ref. 25. These annually resolved δ15N data represent a tracer for water mass changes in the region, where high values are characteristic of the presence of Labrador Slope Water. The time evolution of the δ15N tracer agrees well with that of our AMOC index (Fig. 5). Ref. 25 reports four more data points from ancient corals preceding the twentieth century, the oldest one from AD ∼500. These lie all above 10.5‰ [per mille], providing (albeit limited) evidence that the downward excursion to values below 10‰ between 1975 and 1995 and the corresponding water mass change may be unprecedented in several centuries.
That's right. Rahmstorf15 quite explicitly states that δ15N data represent a tracer for water mass changes.
It looks as if Steve McIntyre was so intent on proving that "scientists don't know nuffin'" that he didn't bother reading the paper. He copied and pasted the Figure 5 caption but didn't take the time to read the section in Rahmstorf15 that described the chart in more detail.
You'll also have noticed that Figure 5 has δ15N data plotted on the right hand axis, not the left hand one (which is a temperature scale for other different data). Maybe Steve missed that, too.
Now I'll show you the chart from Sherwood11, which Steve put up. It's a segment of Figure 3, which is of annual mean bulk δ15N from six colonies of the deep-sea gorgonian P. resedaeformis.
The two are very similar, aren't they. That's because Rahmstorf15 was using the data from Sherwood11. And notice the y axis label. It's not temperature is it. It says Bulk δ15N ‰. It's an indicator of the nutrient status of the ocean in that region over time. The (inverse) correlation with temperature (in that part of the ocean) is through the source of the nutrients - warm subtropical vs cold sub-Arctic.
As I said earlier, what got Steve so excited was that he (mis-)figured that Rahmstorf15 drew the wrong inference and were saying the mean bulk δ15N data represented temperature. Steve crowingly wrote further down:
It’s not even well established that coral d15N is a proxy for local ocean temperature. Coral d15N is not a well-studied proxy, to say the least. Only a few examples are reported with results from only one article being archived at NOAA. In one of the original articles on the proxy, co-authored by the lead author of Rahmstorf’s citation, d15N values were thought to depend on distance from sewage source.
Ha ha. Poor Steve. He doesn't even remember the title of Sherwood13. Yes, naturally enough (pun intended) there are many sources
During negative modes of the NAO, cold (6 °C), fresh (34.9 psu), and nutrient-poor (∼15 μM Graphic) Labrador Slope Water (LSW) is advected along slope, from the Grand Banks of Newfoundland to southwestward of Georges Bank (Fig. 1). During positive modes of the NAO, LSW retreats northeastward and is replaced by relatively warm (10 °C), saline (35.2 psu) and nutrient-rich (∼23 μM Graphic) Warm Slope Water (WSW) associated with the Gulf Stream.
Different regions, different surface temperatures
In case you are wondering, the temperature off the coast of Nova Scotia (Sherwood11) can be quite different to the temperature in the sub-polar gyre region that Rahmstorf15 was about. Here's a segment from a map from NOAA, showing the 2014 temperature as an anomaly from the 1880 to 1920 average. I've marked the different regions (click to enlarge):
It took five whole days before anyone noticed!
A whole five days passed before anyone commented on Steve's big blunder. Most of the commenters were dull and gullible and foolishly took Steve at his word (or what they could understand of his words.) Bill H was the first to point it out, writing much more succinctly than I did (archived here):
Apr 6, 2015 at 12:31 PM
Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature. Your entire post would consequently seem to be based on a false premise.
Steve McIntyre still didn't get it, replying:
Apr 6, 2015 at 3:56 PM
Bill H, I do not understand what exactly you are objecting to. I do not claim to be infallible and try to correct errors when they are pointed out to me. Can you provide a direct quotation from statements that you believe to be erroneous. Simply saying that the entire post is “based on a wrong premise” is arm-waving. It is my understanding that Rahmstorf’s Figure 5 purports to show a positive correlation between Nova Scotia coral d15N and gyre temperature and I have seen nothing in your statements to change this understanding.
No. It's not "arm-waving". Bill rightly pointed out that the entire article is wrong. That it was based on a false premise, not that there was just one little itty bitty mistake. The whole damn thing was wrong. Did Steve really want Bill to quote the entire article?
Steve is stuck on his false notion that Rahmstorf15 included the δ15N data as a proxy for temperature, even though the paper specifically states it's an indicator of water mass changes, writing:
...These annually resolved δ15N data represent a tracer for water mass changes in the region, where high values are characteristic of the presence of Labrador Slope Water.
What follows from Bill H's comments is a lot of silliness, ranging from bristlecones to the disgraced Wegman. Really, I know a lot of deniers are hopeless at science, but surely they know the difference between "water mass changes" and "temperature". (I didn't see a sign of anyone except Bill H recognising Steve McIntyre's big blooper. If there was someone, Bill H might point it out in the comments.)
Update: Bill has pointed out that a commenter called Carrick saw what Bill H saw. I found it. (BTW I first archived this at Archive Today, but it chops off some bits, so the full version is also archived at WebCite, which didn't chop anything - but doesn't provide direct links to different bits of the article. WebCite includes a new comment by Steve - right at the bottom - where he's finally, six days later, showing signs of backtracking but not without some thimble switching.) [Sou - 12:54 am 8 April 2015 AEDT]
Update 2: How Steve's brain is stuck and won't budge
Steve McIntyre still hasn't got it. He's blundering about trying to link the Sherwood11 data to temperature, whereas the scientists linked it to water mass movement. In his latest comment he is still claiming a negative correlation - but it's not. That's because he can't get temperature out of his brain. (Is global warming causing a mind melt?). Well after lots of people attempted to point out his big blunder, Steve's last (latest) comment is:
In my opinion, there is considerable evidence for spatial autocorrelation of temperatures up to 1200 km or so. Indeed, this motivated the discussion of Chladni patterns in connection with our commentary on Steig et al 2009, as principal components applied to spatially autocorrelated data in a finite geometric region yields Chladni patterns.
The average distance from the Nova Scotia corals to the subpolar gyre gridcells used in R15 is 2500 km, double the Hansen radius. In addition, Rahmstorf is not arguing for a positive correlation, but for a NEGATIVE correlation. While NH wave patterns do exist, it’s a large jump to claiming that these are stable enough to yield a permanent negative correlation of temperatures at a distance of 2500 km or so.
And further down he wrote: "nor did Rabbitt support large negative correlations. A complete red herring."
He's wrong. Rahmstorf15 AMOC index showed a positive correlation with mean bulk δ15N as measured by Sherwood11. The AMOC index declined when the mean bulk δ15N declined, and rose when the mean bulk δ15N rose.
Steve's still got his correlations wrong. The positive correlation is with mean bulk δ15N, which is a "tracer for water mass movement". In other words, less Labrador sea water was pouring south into the region where the Sherwood11 coral measurements were taken, over the time that the AMOC was shown to slow. That is, over much of last century and most particularly from around 1970 to 1990. (See the charts above.) That meant that the corals had better access to nutrients. The Labrador slope water is nutrient poor, the water coming up from the subtropics is richer in nutrients. (Lower mean bulk δ15N signifies a more nutrient-rich environment.)
Steve's comment was after more than one person tried to correct his misconception. Even Eli Rabett chimed in with a detailed explanation, and pointed out that teleconnections aren't necessarily about atmospheric connections (they can be about ocean currents, too). Note that the second half of Eli's comment is not Eli, but Steve.
The last paragraph in Rahmstorf15 reads:
Although major uncertainties remain about the past evolution of the AMOC for lack of direct measurements, indirect evidence from multiple sources provides a consistent picture, linking together the time evolution of surface temperature, ocean circulation and, possibly, Greenland ice mass balance. If the interconnections between these three components continue as we have conjectured, the ongoing melting of the GIS, which reached an extreme in 2012 (ref. 44), may lead to further freshening of the subpolar Atlantic in the next few decades. Bamber et al.45 estimate that if current trends continue, the Greenland freshwater input from 1995 to AD 2025 may exceed 10,000 km3. This might lead to further weakening of the AMOC within a decade or two, and possibly even more permanent shutdown of Labrador Sea convection as a result of global warming, as has been predicted by some climate models.
Steve needs to read up on AMOC and Labrador sea convection. Or even look at some pictures, like this one from The Encyclopedia of Earth:
|Topographic map of the Nordic Seas and subpolar basins with schematic circulation of surface currents (solid curves) and deep currents (dashed curves) that form a portion of the Atlantic meridional overturning circulation. Colors of curves indicate approximate temperatures. |
Source: R. Curry, Woods Hole Oceanographic Institution/Science/USGCRP. Linked source: EoE
Or for more detail, this one from the Bedford Institute of Oceanography, which also has a short explanation of why the Labrador Sea is so important:
|Source: Bedford Institute of Oceanography|
Note also that Stefan Rahmstorf has been studying Atlantic circulation probably since before Steve McIntyre had ever heard of AMOC. He had a much cited paper in Nature twenty years ago, back in 1995:
Rahmstorf, Stefan. "Bifurcations of the Atlantic thermohaline circulation in response to changes in the hydrological cycle." (1995): 145-149. doi:10.1038/378145a0 (pdf here)
Who do you "believe" - an expert who has specialised in the subject for decades, or a conked out denier blogger with no scientific training or expertise, who can't tell the difference between temperature and mass movement? That's a rhetorical question :D
Added by Sou 8 April 3:40 pm AEDT
Rahmstorf, S., J.E. Box, G. Feulner, M.E. Mann, A. Robinson, S. Rutherford and E.J. Schaffernicht, "Exceptional twentieth-century slowdown in Atlantic Ocean overturning circulation", Nature Climate Change, 23 MARCH 2015 | DOI: 10.1038/NCLIMATE2554
Sherwood, Owen A., Moritz F. Lehmann, Carsten J. Schubert, David B. Scott, and Matthew D. McCarthy. "Nutrient regime shift in the western North Atlantic indicated by compound-specific δ15N of deep-sea gorgonian corals." Proceedings of the National Academy of Sciences 108, no. 3 (2011): 1011-1015. doi: 10.1073/pnas.1004904108 (open access)