Ocean Acidification and Corals

Richard S Courtney (07:08:15)
You continue to make assertions that are not only unsupported by evidence, but are demonstrably untrue. Let’s look at your comments on CO2 levels, and particularly carbon isotopes, in the light of the data that is accessible to policymakers and their scientific advisors:
The 13C data can be obtained by direct measurement (Mauna Loa since 1959; numerous other measuring stations) and from high resolution ice cores (especially the Law Dome cores). Some of the data and the methodologies are described here, for example:
Francey RJ, Allison CE, Etheridge DM, et al. (1999)
A 1000-year high precision record of delta C-13 in atmospheric CO2 TELLUS B-Chem Phys. Meteor 51, 170-193.
Meure CM et al. (2006) Law Dome CO2, CH4 and N2O ice core records extended to 2000 years BP Geophys. Res. Lett. 33, art # L14810
D. M. Etheridge et al (1996) “Natural and anthropogenic changes in atmospheric CO2 over the last 1000 years from air in Antarctic ice and firn. J. Geophys Res. 101, 4115 -4128
Here are some of your assertions:

4. The change to the 12C:13C isotope ratio of atmospheric CO2 is in the direction expected if the recent increase in atmospheric CO2 concentration were caused by the anthropogenic emission of CO2.
But if the ratio changes then there is a 50:50 chance that it will change in that direction or the other.

That doesn’t work. The 12C:13C ratio isn’t just “in the direction expected” due to “a 50:50 chance that it will change in that direction or the other”. Inspection of the data demonstrates that the change in 12C:13C ratio has been in the “expected direction” continuously throughout the last 300 years or more. When our emissions were very low, the change in 12C:13C was very low (but in the “expected direction”). When our emissions were middling, so was the change in 12C:13C (still in the “expected direction”!), when our emissions were large, the change in the 12C:13C ratio was/is large (still in the “right direction”).
So for example we can inspect the record (see data in Francey cited above, for example) and find that the annual changes in delta 13C are negative (as carbon is returned to the atmosphere from burning fossil fuels the 13C content of atmospheric CO2 <reduces since fossil fuels are 13C-depleted). Here’s some of the data grouped into periods. The change in 13C:12C (- delta, delta 13C per year) was not much above zero in the period 1700-1800, started to rise through the 19th century, was rather substantial during the first half of the 20th century, and rose very quickly in the 2nd half of the 20th century as our emissions similarly rocketed upwards during this period. Now the -delta,delta-13C values are changing by around 25e-2 per year.
year range…………..(-)delta-delta-13c
1700-1800………….<0.3e-3 per year
1800-1850………….1.5-2.1e-3 per year
1850-1900………….3.3e-3 per year
1900-1950………….6.2e-3 per year
1950-2000………….~18e-2 per year

5. The magnitude of the change to the 12C:13C isotope ratio of atmospheric CO2 is much smaller than expected if the recent increase in atmospheric CO2 concentration were caused by the anthropogenic emission of CO2.

That’s not true. If anything the rate of change in 13C:12C is a bit faster than expected in recent decades since the rate of increase of atmospheric CO2 is faster than the 13C-replenished atmospheric CO2 is able to equilibrate with CO2 in the surface layers of the oceans. See Figure 11 of Francey et al cited above.

6. The fact in point (5) indicates that …..

Since the premises are false [your point (5) doesn’t contain a fact], your conclusions from here on are erroneous…

Phil,
Given that there are quite a few years in the graph with “calculated data” that had no “measured data,” it is clear that the “calculated data” is not derived from any reported measured data.
If you download their measured data set, you can see that there are many years with no data and that the data set corresponds with the red (measured) data in the graph.
So far people have reported three raw measured pH data sets. Monterey, Aloha and Kahe. None of them show a statistically significant trend towards lower pH.

J. Peden (09:51:57)
n.b. in order to be crystal clear regarding the meaning of “this” as per the statement below from the above-referenced post:
that is, when you have not shown this to be the case anywhere else
“this” refers to “that CO3 must decrease when pH decreases from the addition of CO2 to solution”

Graeme Rodaughan (04:43:05) :
Simon Evans (08:41:06) :
Graeme Rodaughan (07:58:45) :
In this forum – I also raise my hand and second Alan Millar’s request.
Alan Millar’s question was addressed as follows:
Perhaps some of our alarmist friends would like to explain.
I’m not an alarmist, so I wont be responding to that.
Fair enough – I’ll re-ask politely.
Is anyone able to explain how a warming ocean could absorb CO2 and become more acidic – given that warming oceans have a well known property of outgassing CO2? – Or is this simply a non-starter?
Thanks.
Thanks Graeme. My explanation would be that both occur at the same time – outgassing and absorption, that is – the net effect simply being a matter of whichever is the greater, according to the variables of temperature (and also pH, a lowering of which will itself reduce the sink capacity of the oceans) and atmospheric concentration. At the moment the oceans are absorbing more than they are emitting, but the projections are for the reverse. Therefore, the extent of acidification is limited and, eventually it would be undone. The concern for the corals is that they will be destroyed in the interim. I should stress, though, that this is not my field, so don’t take my word for it! 🙂

J. Peden:
Btw, where do you think CO3 comes from to begin with?
CaCO3! We’ve been beating that drum for the last 100 posts!
You say that the Henderson Hasselbach equation is too simple to definitively show that CO2 levels reduce CO3(-2) equations. But this complicated science has ALREADY been done AND observed.
I’ll quote myself again:
http://www.publicaffairs.noaa.gov/releases2002/mar02/noaa02r305.html
“Our results are important in understanding the ocean’s role in the global carbon cycle,” Anderson said. “Prior to this study, large changes in ocean carbonate chemistry had been proposed to explain the changes in atmospheric carbon dioxide. Over thousands of years, calcium carbonate compensation appears to be the dominant variable controlling the ocean carbonate (and carbon dioxide) inventory. When carbon dioxide from the atmosphere is added to the oceans, the calcium carbonate on the seafloor dissolves to minimize the carbon dioxide change in the ocean. ”
If [CO2] increases [CO3(-2)] in the ocean, shouldn’t it be easier to form CaCO3 due to this equilibria: CaCO3 == Ca(++) & CO3(-2) according to Le Chatlier’s principal? In fact, the opposite is shown to be true. The addition of H+ ions is more important to [CO3(-2)] than the dissociation of HCO3(-1) by CO2.
You guys are too focused on the generation of CO3(-2) by CO2 in pure water. CaCO3 is the much greater buffer (which is why oceans are alkaline and not acidic).
I rest my case again.

Hmmmm … still no takers.
OK … 2 sigma.
(Meaning, show me, that the “signal” of lowering pH can be demonstrated, via a classical Gage R&Red set of runs, to be able to demonstrate a 2 sigma margin versus all errors and noise vis a vis repeatability and reliability, plus, bias factors, across a semi infinite half space of ocean mass).

“” foinavon (13:02:16) :
George E. Smith (12:07:00)
George, we’ve already had the dismal non-”argument” about the meaning of “acidification”. Are you really going to proceed down a pedantic diversion over the meaning of “sinusoidal”?
Remember that words are only labels for the things that we are choosing to describe.
“Sinusoidal” is actually not a bad description of the very well-characterised cyclic variation in the atmospheric CO2 concentration that follows the seasonal growth/decay cycle in the N. hemisphere which is where most of the earth’s flora “resides”. Since the sinusoidal variation lies on a rising trend of growing atmospheric CO2 it does look a bit “sawtooth”.
So let’s call it “sawtooth” if you like. We both know what we’re looking at. “”
Well Foinavon; contrary to what you appear to have assumed, my comment was not at all a “pedantic diversion”; and it was not intended to be dismissive of your comment. If you took it that way, then I apologise.
BUT; would you not agree that the distinction between “roughly sinusoidal” and “roughly sawtooth”, is very significant in the sense that each would be the result of quite different physical processes.
For example, when I look at the satelite graphs of the solar constant measured over sunspot cycles, the resulting variation I would call “roughly sinusoidal”, or even “sinusoidal” if trying to be brief; yet it clearly isn’t sinusoidal, because the peaks are sharper than the troughs, indicating the presence of a second harmonic component. That then begs the question of what physical processes going on in the sun lead to that second harmonic deviation from a sinusoid.
So back to the ML data. The peak and trough turnarounds are very sharp transitions from upslope to downslope and vice versa. Clearly this is not the result of a nearly linear feedback oscillator, which would produce a sinusoidal oscillation.
There is a definite on/off transition occurring, which cause the ML cyclic seasonal variation; and that leads to the question; what is that on/off mechanism.
Now for one thing, deciduous trees lie dormant through the fall/.winter, and then sprout leaves (very quickly) in the spring. I have a home in California’s central valley agricultural area where stone fruit orchards abound (and grapes). Those plants are sprouting leaves en masse, as I type this.
Those trees in the winter, are not taking up CO2 since they have no leaves. The leaves return in the spring in a matter of a week or two; and from there they start sucking up CO2. To a large extent, the mass of foliage doesn’t change after that two week leaf growth period, so you have a rather constant CO2 extraction rate once the leaves are there, and such a mechanism would in fact lead to a saw tooth cyclic variation in CO2. To me the leaf growth/shedding processes of spring and fall, amount to a rectangular pulsed waveform, with rise and fall times of the order of two weeks, but since the time constant of the excess CO2 uptake by that growth process is long comared to the season, the CO2 response is an integration of that rectangular drive signal.
NOW ! I am NOT claiming that this is the process which explains the ML cyclic waveform; but I am saying it is one operating process, which would create such a response.
If you find, and look at the NOAA three dimensional CO2 graph form pole to pole, you will notice that the southern hemisphere cycling is much suppressed from the northern, and to my eye, the very limited south pole cycling, is in fact much closer to a “roughly sinusoidal” variation (still with harmonic content) than it is to a saw tooth; and that suggests to me that the southern CO2 uptake, is likely more ocean related, than land plant growth.
So you see, I was not trying to be smart alecky; but hinting that the waveform isa clue to the mechanism.
But as I said, if you took it as frivolous comment on your post please accept my apology; it wasn’t intended that way at all.
George

@ur momisugly Steven Goddard et al.,
There still seems to be a great deal of confusion over the meaning of ‘calculated’ vs. ‘measured’ for pH data from the HOT datasets, despite that fact that I’ve explained the meanings at least twice now. In any event, let me try to communicate once and for all how these data were generated.
Seawater pH can be assessed using three methods:
1) With combination electrodes calibrated in appropriate buffers (NIST buffers don’t work well—one needs to use Tris-based buffers made in an artificial seawater solution).
Of the three methods available, this is the LEAST ACCURATE method, and wasn’t used to obtain any of the HOT (or BATS, etc.) data. While this method can be precise to better than +/- 0.005 pH units, it is rarely accurate to +/- 0.005 pH units or better.
2) Spectrophotometrically using the indicator dye m-cresol purple (Clayton and Byrne, 1992). This method did not exist when people started taking the HOT data, therefore could not be used.
Accuracy, as has been assessed in the lab and on many NOAA cruises, is on the order of +/- 0.001 pH units or better. Precision (assessed likewise) is typically +/- 0.0004 pH units or better, and can be up to an order of magnitude better. This is a good method to determine pH when very high accuracy is needed, but unfortunately is somewhat labor intensive (I know, I’ve used this method extensively). This method was used to determine pH in SOME of the Niskin water samples from these sites from 1992-1997, and 2003-present.
The ‘measured’ pH data for the HOT datasets were assessed using this method.
3) Calculated given measured values of:
1) total CO2/DIC, 2) total alkalinity, 3) salinity, 4) total P, 5) total Si, 6) in situ pressure, 7) in situ temperature, 8) experimentally determined dissociation constants (e.g., K1, K2) for carbonic acid.
The accuracy and precision of this method are similar to the m-cresol purple spec. method (~accuracy +/- 0.0007 or better, precision +/- 0.0005 or better).
The ‘calculated’ pH data for the HOT datasets were assessed using this method, and can be re-assessed ad nauseum. All of these data are freely available on the website, and we can produce our own analyses of these results all we want all day long everyday. These data (along with a great deal more) cover the entire period.
Something worth pointing out is that all of these methods, and indeed EVERY possible method of determining pH relies on calculations from measured values. All provide a ‘calculated’ pH relative to experimentally determined calibration values.
The ‘measured’ data in the graphic I presented are for a given subset of the Nisken bottle samples during the sampling period. Only a subset of the Nisken samples could be analyzed because these measurements are somewhat time sensitive, so have to be analyzed rather quickly after sampling.
The ‘calculated’ data in the graphic include data from a much larger set of Nisken bottle samples.
Hence, these data were collected at the same general sites, but they represent measurements of different water masses. The ‘calculated’ values are broader in scope while the ‘measured’ values include only a subset of the same water masses.
If I may use an analogy, the ‘measured’ values here are like measuring the rate of unemployment in all cities in Florida with a population greater than 70,000. The ‘calculated’ values here are like measuring the rate of unemployment in all cities in Florida with a population greater than 5,000. They will inherently include some of the same values and express the same property, but not over precisely the same landscape.
If you still don’t understand where these data come from (and keep in mind, we can analyze and re-analyze, and re-re-analyze all we want since all of the data are freely available) I’m not sure how much more I can help. If you’d like help in obtaining the pertinent data and in analysis I’m happy to get you started.
Claiming that the data should be thrown out simply because you don’t understand where they came from (it’s all in the references/methodology), however, is ridiculous, and terribly nieve. Again, your ignorance is not evidence.
Best,
Chris
p.s. And beyond this point of contention, look at the big picture: if CO2 is added to the atmosphere, is it logically possible that some of it will not dissolved into the ocean? (no, of course not) Is it logically possible (notwithstanding J. Peden et al.,’s severe misunderstanding of carbonate chemistry) to add CO2 to sea water and not reduce the pH of that water? (no, of course not)

Chris J,
If you look at the “calculated” pH data, it has an extremely high standard deviation, does not correlate well with the measured pH data in coincident years, and shows a large inexplicable drop in pH from 1989-1994 (before the measured pH data set started.) That drop, plus inconsistent low calculated readings the last couple of years accounts for almost their entire claimed trend. The calculated data shows variations as high as 0.8 during time periods when the measured data has a peak variance of only 0.2.
I am surprised that they would include such an obviously flawed data set in their trend analysis.