EM

“There is also a 0.2 downward step around the 1998 El Nino, too small to be significant but in the opposite direction to what you expected.”

Did you read what I said previously? Try again: When an El Niño kicks in, the δ13C drops more rapidly. Precisely in the direction I stated. BTW, how on earth would you know what I “expected”. I don’t approach data analysis with expectations, I listen to what the data are telling me.

“Both are small enough to be noise, rather than causation.”

If you look at only one location, such as the South Pole, you can possibly argue this point. However, when you see the same variations at the same time at different locations (i.e. from a different set of CO2 samples taken thousands of miles away) the probability that these changes in rate of decline reflect random noise diminishes rapidly. Compare the South Pole data with the data from Mauna Loa:

http://scrippsco2.ucsd.edu/graphics_gallery/isotopic_data/mauna_loa_and_south_pole_isotopic_c13_ratio

Unfortunately, these plots are not seasonally corrected, but it is still possible to identify the El Niños by the increase in rate of decline for 82-83, 86-88 and 97-98. The El Niño of 91-92 is not seen, presumably being masked by the effects of Pinatubo. The plots I prefer to use are self-generated from the Scripps’ seasonally corrected files. They are saved on TinyPic, but if I try to link them from here, you are directed to their horrible website with dubious adverts everywhere. On other blogs I am able to simply link to the graph itself. If anyone knows how to get around this issue, please let me know.

“The long term decline is due to increasing 12CO2 in the atmosphere.”

The long term decline in δ13C in atmospheric CO2 is due to the incremental CO2 having, on average, a lower 13C/12C ratio than the current atmospheric level (where the δ13C is around -8.5 per mil). In fact, the climate science position (if I can call it that) is that having made the assumption that all the CO2 growth is anthropogenic with a δ13C content of circa -28 per mil, the modelled decline rate is much too high and it is necessary to “add back” very substantial amounts of 13C to the atmosphere from the oceans in order to get anywhere near a match! Bear with me on this point.

We can use the so-called Keeling plot (plot 1/CO2 vs δ13C) to estimate the 13C/12C ratio of the incremental CO2. It is -13 per mil (and has been since 1750 or thereabouts if you accept the Law Dome ice core data). This value is not controversial in itself, as far as I know, but is certainly not broadcast widely in the climate science literature. How often have you heard that the δ13C decline rate is consistent with an anthropogenic source? Directionally it is, but not in absolute terms. A myth!

The best explanation of the climate science position on δ13C that I have seen is here:

www.onlinelibrary.wiley.com/doi/10.1029/2001GB001845/pdf

See, in particular, Fig 5. As it states, the brown vector represents estimated fossil fuel emissions. Since 12C is 99% of CO2, we can use the gradient of this vector to check the δ13C value. It is -168/6, i.e. -28 per mil (as stated in Table 1 of the paper). The black vector “represents the observed change in atmospheric composition”. So it has δ13C of -38/3, i.e. -12.67 per mil, close enough to -13 per mil, I think you will agree. The dark green and dark blue vectors represent the assumed uptake of CO2 by land and ocean sinks respectively (totalling half of the emissions in terms of CO2), but providing only a partial adjustment of 13C flux and still a long way from the known atmospheric content. The final adjustment, which corresponds to 33% of the total 13C flux from emissions, is the “ocean disequilibrium forcing” which provides the additional 13C that I mentioned above. This is based on the assumption that CO2 exchange between the oceans and the atmosphere leads to a net increase in atmospheric δ13C even without any net effect on CO2 levels.

This an interesting model, made even more complicated by the view that isotopic discrimination may be varying with time. My problem is this: how does this model fit with the incremental CO2 having a constant long term net δ13C content from 1750 to the present day?

" I am using "significant" in the scientific sense of statistical confidence."

What is the exact meaning of that?

I have always used my own bodged 'commonsense' view of statistics. If you can't actually display your data in way way where you can visually see the explanation of it and have to draw a straight line through a massive crowd of dots with a 'statistical' straight line through them I will always really doubt the conclusions.

May 1, 2018 at 12:23 AM | Rob Burton

The significance of significance lies with the confirmation bias of the Climate Scientists and how they are beholden to their funding.

https://en.wikipedia.org/wiki/Statistical_significance

May 1, 2018 at 3:21 PM | Phil Clarke

How has Climate Science defined the null hypothesis, to claim anything significant has happened that has not happened before?
"The null hypothesis is the default assumption that nothing happened or changed"

"The null hypothesis is the default assumption that nothing happened or changed"

This definition reminded me of the δ18O-CO2 data. So how do you rate the significance of this dataset? Is the fact that it has not changed in 35+ years "significant" or not. It even manages to maintain the same offset between observatories. How do you (EM, Phil, anyone) explain the observations?

http://scrippsco2.ucsd.edu/graphics_gallery/isotopic_data/global_stations_isotopic_o18_trends

I am genuinely interested because there is very little in the literature that I can find that addresses this point. A possible clue: it seems that there is oxygen isotopic interchange when CO2 is in contact with H2O (which led to some early δ18O-CO2 measurement issues due to undried flasks).

Golf Charlie

For AGW the null hypothesis is that the climate is being determined by natural variation, with no human influence.

You will remember the six temperature control knobs; solar insolation, orbital cycles, plate tectonics, albedo, land use and CO2.

Measure them and you find that plate tectonics is set to icehouse, orbital cycles are set to interglacial, solar insolation, land use and albedo are set to neutral and CO2 is set to rapid warming. At their current settings he sum of the five natural forcings is zero and constant, so the null hypothesis is indeed no change.In this case the null hypothesis is that under natural conditions the climate system would have 280ppm CO2 and a global average temperature of 13.8C.

Since CO2 is the knob we are controlling, the alternative hypothesis is that our industrial emissions are causing a difference between observed temperatures and the null hypothesis consistent with the effect of increasing CO2.

How to test? The CO2 and temperature are both significantly larger than the null values.

Is the warming due to CO2? The observed warming matches projections from theory and laboratory scale experiments. To be sure would need controlled trials using duplicate planets and several centuries.

What a pity that the last two years have seen the biggest 2-year drop of global temperatures yet recorded, of over 0.5°C. 

Temperatures fell more than 0.8C in the 12 months starting Jan 2007.

HADCRUT4 Global mean

There was a similar drop following the 1998 El Nino. Meanwhile the 30 year trend is warming at the rate predicted by models, that include greenhouse gas forcing.

JR

When CO2 dissolves in water you get carbonic acid, then bicarbonate ions in anequilibrium reaction.

CO2 + H2O <> H2CO3

H2CO3 <> HCO3- + H+

That is why you get a pH decrease with more dissolved CO2 and is the reason for the oxygen isotope interchange you mentioned.

18O will also be swapped back and forth between water, CO2 and O2 during combustion, photosynthesis and respiration. You can probably regard the 18O in all three as a single reservoir.

There is an 18O paleo temperature test used in ice cores.
Heavier H2 18O evaporates less readily from ocean surfaces than H2 16O. Thus ice cores are depleted in 18O relative to the oceans. Since the depletion effect is less in warmer water, more 18O in an ice core indicates a warmer global temperature.

The Point

Continued

The Point Barrow data shows a maximum in Spring. I wonder why.

Offhand I can see no reason why global warming would produce a long term change in the proportion of 18O in the whole system, though you might find a bit more in the ice as discussed.

May 1, 2018 at 8:04 PM | Unregistered CommenterEntropic man

" I am using "significant" in the scientific sense of statistical confidence."

Again, a very simple question with simple answer. As a science expert what is your basic definition 'exactly' of statistical significance/confidence.

RobBurton

Two means may be regarded as significantly different when the probability of that difference occuring by chance is less than 5%.

At that point a scientist will be confident that there is a reason for the difference, and go looking for the explaination.

I think I have worked out why the Northern Hemisphere shows a seasonal cycle for 18O in CO2.

During the Northern Hemisphere Spring and Summer there is a higher proportion of 18O in water evaporating from the oceans and precipitating onto land.

This is taken up by plants, used in photosynthesis and built into their structure.

When they decompose during Autumn and Winter the 18O is released in CO2, increasing the proportion of 18O in atmospheric CO2 during the Winter.

EM

So what is the probability that the match of δ13C trend variations between Mauna Loa and the South Pole is purely by chance?

http://scrippsco2.ucsd.edu/graphics_gallery/isotopic_data/mauna_loa_and_south_pole_isotopic_c13_ratio

EM

Interesting hypothesis re δ18O. What we do know for certain is that there is no reflection of the growth in atmospheric CO2 within the atmospheric δ18O-CO2 observations.

I hadn’t looked in detail at the δ18O-CO2 annual cycle previously. A quick look at the Point Barrow data reveals that it generally matches the character of the atmospheric CO2 cycle quite well with a flattish max in many cases during the winter and a sharp minimum in (very) late summer. However, it also shows a consistent lag of around two months: the δ18O cycle lags the CO2 cycle. A reflection of an equilibration process?

There is evidence for another (longer) cycle in the δ18O-CO2 observations for which I can find nothing in the literature. There is probably not yet enough data to conclude anything definitive, but see here:

https://www.esrl.noaa.gov/gmd/dv/iadv/graph.php?code=MLO&program=ccgg&type=ts

Select the parameter: Oxygen-18/Oxygen-16 in Carbon Dioxide. The two cycles can also be seen in the Point Barrow data on the same site. So we have at least two indpendent sources of the tends. The only paper I have found that relates to this observation was published in 2002 and it tries to explain the downward trend during the 90s, which has subsequently been shown to just part of a cycle!

One more thing. I have yet to see an explanation for the atmospheric δ18O-CO2 decreasing from the South Pole (circa +1.0 per mil) to Alert in the north (-1.0 per mil). Any suggestions?

JR

All the biologically driven cycles show much more strongly in the Northern Hemisphere because most of the land and most of the biomass is in the Northern Hemisphere.

Hawaii spend at least half the year North of the ITCZ, in the Northern hemisphere airmass, hence the visible seasonal cycle.

The Southern Hemisphere and the South Pole would be expected to show a cycle 6 months out of phase, but much less strongly because the SH has much less land.

For arithmetical reasons the average for a cycle would be slightly smaller than the same average without a cycle. I doubt that there is any real difference in the 13C concentrations in the two hemispheres.

EM

I have not made any reference to the size of the seasonal CO2 cycle in my comments thus far so I am not sure why you have raised this subject. My comments about the δ13C variations were expressly in relation to the changes in trend that were distinct from the seasonal cycle. Why else would I be using the seasonally corrected data! Still, since you have raised the issue of the size of the seasonal CO2 cycle, here are my views:

The annual CO2 cycle is dominated by the effect of the boreal forests. This is why the onset of the reduction in atmospheric CO2 at Point Barrow is in May (“Tim Lueker, research scientist in the Scripps CO2 Research Group, only needs one sentence to explain why atmospheric CO2 peaks in May. “Springtime comes in May in Siberia,” he says”). It is also the reason for the annual cycle at Alert being slightly smaller than at Point Barrow (it’s a long way from the forested areas). Boreal forests are absent from the southern hemisphere, so only a very small annual cycle is seen there. But what about the equatorial rain forests of the Amazon, central West Africa and South East Asia? The reason that we do not see much of a seasonal cycle there is because photosynthesis is temperature dependent and in these areas there is very little annual temperature variation (unlike the boreal forests). So, plenty of CO2 exchange is undoubtedly going on but at roughly constant rates throughout the year. This should be evident in the diel cycle.

See here for forest distribution: https://askabiologist.asu.edu/sites/default/files/resources/articles/biomes/world-biomes-map-540.gif

You say: “I doubt that there is any real difference in the 13C concentrations in the two hemispheres.” However, the question I asked of you was with respect to atmospheric δ18O-CO2. Do you think that the latitudinal offset of δ18O from +1 per mil to -1 per mil is real?

18O and 13C annual cycles that involve biological activity will all synch to the seasons in the Northern Hemisphere because that is where most of the biomass is.

"Do you think that the latitudinal offset of δ18O from +1 per mil to -1 per mil is real?"

No

A straightforward answer – good! However, as a scientist with expertise in statistics, you will surely appreciate that the probability that the latitudinal trend, as supported by observations at 11 different locations, is occurring by chance would be miniscule. If you do not accept the data, there is nothing left to discuss.

JR

For a valid scientific theory you need observed data and you need a mechanism.

A gradient is generated by a source at the high end and a sink at the low end.

What would you suggest are acting as source and sink?

Your idea of how science works and mine are rather different. I start with the data. Are the trends/variations meaningful in the light of the accuracy and precision of the data? For this, you do not need much in the way of specialist (climate science) expertise, only reasonable data analysis capability. This is where I am. If yes, i.e. the trends are meaningful, then we can move on to consider hypotheses. These hypotheses have to be tested against other evidence before they can possibly be considered as a theory. In my view, there are no theories in this area, i.e. in regard to the explanation of the δ18O-CO2 data. The published hypotheses struggle to withstand the rigours of testing.

The reason for the constant offset in δ18O-CO2 with latitude is, for me, fascinating. It suggests some kind of equilibrium with the oceans and/or biosphere rather than an issue of sources/sinks.

Although it is a different gas, I believe that the methane data are extremely important in understanding this offset with latitude because the offsets in CH4 are clearly beyond the uncertainty in the data (a mean value of approx. 1,800 nmol/mol at the South Pole and 1,950 nmol/mol at Point Barrow ) and yet changes in growth rate occur at the same time despite this. Have you studied the CH4 data?

Sometimes science can be fun.

https://pubs.geoscienceworld.org/ssa/srl/article-abstract/525827/do-large-magnitude-8-global-earthquakes-occur-on

You aren't looking at a linear trend.

ALT and PTB are showing -1.0

NZD and SPO are showing 1.0.

That other seven sites are all showing 0.5.

You have a flat graph with a lift in high Southern latitudes and a drop in high Northern latitudes.