US - Source or Sink?
Coby Beck
due to terrestrial uptake. I have responded to that on my blog here:
http://illconsidered.blogspot.com/2006/04/us-is-net-co2-sink.html
It all seems pretty straightforward.
On Jennifer Marohasy's aussie blog I have presented that page in a
thread here:
http://www.jennifermarohasy.com/blog/archives/001383.html
Now someone has provided this abstract:
http://www.sciencemag.org/cgi/content/abstract/282/5388/442
which has found NA has "a mean uptake of 1.7 ± 0.5 Pg C/year."
My sources say "U.S. emissions of greenhouse gases in 2004 totaled
7,122.1 million metric tons carbon dioxide equivalent (MMTCO2e)"
ftp://ftp.eia.doe.gov/pub/oiaf/1605/cdrom/pdf/ggrpt/057304.pdf
Now, 1 Pg = 1,000 million metric tonnes so we are in the ballpark of
17% US emissions offset by land sequestration but the conversion from
CO2 to just C blows that away.
What am I doing wrong or misuderstanding?
Coby
Tom Rees
So the science abstract says N Am emissions are 1.6 pg carbon, = 5.9 pg
CO2 . This is a bit less than 7.2, and the difference is probably due
to non CO2 GHG emissions and Canadian emissions.
According to the paper, N Am absorbtion is 1.7 pg and emission is 1.6.
Brian Jackson
the bottom of the Fan et al. abstract you posted. In particular, there
are a couple of technical comments by Holland et al and Kaminski et al
(non-subscription) that are illuminating. The impression I get is that
Fan et al results are (i) only for the years 1988-92, where the update
is highly variable and so application to other years may be
unwarranted, (ii) highly uncertain, since the inverse modelling
approach used is based on sparse data, and (iii) obtains a much larger
sink than other estimates, including other inverse modelling studies
and direct estimates such as the one you looked at.
There's also another paper by Pacala et al (with some of the authors of
the Fan et al paper, as well as others) in Science 292, 5525 (2316)
which also looks to have found lower numbers for the U.S. sink
(0.3-0.58 PgC/year).
Hopefully somebody with more knowledge than me can answer, since mine
is entirely based on reading the links below the abstract. But it seems
to me that it may be a case of just considering one outlier paper, and
ignoring all the rest.
Brian Jackson
> There's also another paper by Pacala et al (with some of the authors of
> the Fan et al paper, as well as others) in Science 292, 5525 (2316)
Sorry, that should be Science 292, 2316 (2001)...
Coby Beck
Thanks Brian. Someone following Jennifer's blog forwarded a couple of
the papers you mention to me, and it seems you're right this paper is
just an outlier and time is not treating it well either. I was
concerned I had just namgled the conversion from Pg carbon to MMTCO2e
Coby
Tom Rees
http://www.pnas.org/cgi/content/abstract/98/26/14784 They estimate the
US Carbon sink to be 0.14153 Gt/C/yr for the period 1981-1999. Isn't a
Gt the same as a Pg? In which case their estimate is one tenth that of
Fan et al. Russia is the big winner in the sequestration stakes. Myeni
et al briefly discuss Fan et al too.
Also of interest: The missing Carbon sink:
http://www.whrc.org/carbon/missingc.htm
Coby Beck
>
> Also of interest: The missing Carbon sink:
> http://www.whrc.org/carbon/missingc.htm
Thanks, that's a nice link.
Coby
Alexi Tekhasski
Coby, on your blog
http://illconsidered.blogspot.com/2006/03/natural-emissions-dwarf-humans.html
I have raised a question regarding accuracies in estimation of global
CO2 fluxes, which went unanswered. Maybe this group can help us? I
wrote:
<i> Regarding my concern about accuracy of global CO2 fluxes:
As I gathered the theory behind the estimations, one need:
(1) reconstruct the whole SST field from limited number of met stations
or satellite views. The reconstruction carries errors, satellite
reconstructions use corrections for cloud cover and ocean surface
albedo that includes assumptions and uncertainties;
(2) reconstruct the atmospheric wind field over the same area [how do
they do it?];
(3) apply experimentally-determined "piston velocities" to determine
local CO2 fluxes;
(4) average the product over whole sea area, and over the whole year.
The fields of Co2 concentration, wind velocity, and SST need to be
evaluated at least on monthly time scale in order to get correct
estimation of their product since the variables fluctuate over space
and time.
(some very useful and fascinating details can be found in this CO2
LECTURE.)
http://stuff.mit.edu/afs/athena/course/12/12.842/www/co2lecture2.ppt
The measured "piston velocities" (see slides 16 and 19) have error bars
and general scatter of approximately +-80%.
Now, could you tell me how it is possible to esimate global CO2 uptakes
and inhales at about 100Gt/y with the above product of uncertainties,
yet to arrive at a small difference (ocean uptake) of 2GT/y with an
alledged error margin of only +-0.8Gt/y? Keep in mind that +-0.8Gt/y
margins require estimations of two global fluxes with 0.4% accuracy,
while piston velocities have 80% error margins alone. What did I
miss?</i>
==== end quote
So, regarding your current topic, how do you think it is possible to
determine whether a small difference between two big (and poorly known)
numbers is a source or sink, then claim a number with 5-digit
precision, and seriously discuss any consequences?
- Alexi
EliRab...@yahoo.com
Let me take a stab at this:
For the same reason that the average of a distribution is known more
precisely than the standard deviation. Your limits require that all
errors go in a single direction. Fat chance (actually why don't you do
the calculation). James may have something to baysian about this.
Eli (a Rabett in exile)
Michael Tobis
> Now, could you tell me how it is possible to esimate global CO2 uptakes
> and inhales at about 100Gt/y with the above product of uncertainties,
> yet to arrive at a small difference (ocean uptake) of 2GT/y with an
> alledged error margin of only +-0.8Gt/y? Keep in mind that +-0.8Gt/y
> margins require estimations of two global fluxes with 0.4% accuracy,
> while piston velocities have 80% error margins alone. What did I
> miss?
It is often the case that global quantities are better constrained
than regional ones.
Consider the case of net precipitation/evaporation. Globally, it must
be very near zero on appropriate time scales (say, annual); in any
subcontinental region the uncertainty may be large.
A similar situation applies to carbon fluxes. They are globally much
better constrained than they are locally. Although, alas, not
constrained to be near zero on annual or millenial time scales, they
are constrained by measurements of concentrations and isotope ratios.
The water cycle and the carbon cycle are analogous in that the
horizontal fluxes between geographic regions are likely to be large
and difficult to estimate, but these must globally sum to exactly zero
always.The system is tightly coupled in ways that are not expressed in
regional analyses.
This is not a complete answer to how well constrained the main
reservoirs actually are, but it does answer your "how is it possible"
approach. Sometimes the best way at the whole is not to sum the parts,
mt
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > Now, could you tell me how it is possible to esimate global CO2 uptakes
> > and inhales at about 100Gt/y with the above product of uncertainties,
> > yet to arrive at a small difference (ocean uptake) of 2GT/y with an
> > alledged error margin of only +-0.8Gt/y? Keep in mind that +-0.8Gt/y
> > margins require estimations of two global fluxes with 0.4% accuracy,
> > while piston velocities have 80% error margins alone. What did I
> > miss?
> >
> > So, regarding your current topic, how do you think it is possible to
> > determine whether a small difference between two big (and poorly known)
> > numbers is a source or sink, then claim a number with 5-digit
> > precision, and seriously discuss any consequences?
> \
> Let me take a stab at this:
>
> For the same reason that the average of a distribution is known more
> precisely than the standard deviation. Your limits require that all
> errors go in a single direction. Fat chance (actually why don't you do
> the calculation). James may have something to baysian about this.
>
Which distribution is known?
As I understand, the "distribution" of CO2 concentrations in
oceans was collected over 8 years, with cross-checking for
consitency at 1500m below the surface, that's true.
However, my question was about the global air-surface
exchange. I cannot find any mentioning of direct measurements
of CO2 fluxes across the surface.
To measure it, one needs to measure (a) surface concentrations,
(b) simultaneous surface temperature, (c) simultaneous wind
speed, (d) air temperature (e) pressure?...
and then the "stagnant film thickness" coefficient has to be
factored in. Therefore, all measured data are multiplied by the
_same_ coefficient known to no better than +-80% (in my
eye-ball estimation of the piston velocity graph). Therefore,
the average of the "distribution" must also have at least
the same uncertainty.
Also, I don't see the wind data in WOCE database, and CO2
concentration measurements usually start from 195m deep.
What am I missing again?
- Alexi
Alexi Tekhasski
> On 6/1/06, Alexi Tekhasski <apredt...@austin.rr.com> wrote:
>
> > Now, could you tell me how it is possible to esimate global CO2 uptakes
> > and inhales at about 100Gt/y with the above product of uncertainties,
> > yet to arrive at a small difference (ocean uptake) of 2GT/y with an
> > alledged error margin of only +-0.8Gt/y? Keep in mind that +-0.8Gt/y
> > margins require estimations of two global fluxes with 0.4% accuracy,
> > while piston velocities have 80% error margins alone. What did I
> > miss?
>
> It is often the case that global quantities are better constrained
> than regional ones.
>
> Consider the case of net precipitation/evaporation. Globally, it must
> be very near zero on appropriate time scales (say, annual); in any
> subcontinental region the uncertainty may be large.
Theoretically, it must be near zero, ok. However, what guarantees
the measurement of it to be near zero if you have only a few
hundred measuring points, and each one has the same
bad inaccuracy of 80%?
> A similar situation applies to carbon fluxes. They are globally much
> better constrained than they are locally. Although, alas, not
> constrained to be near zero on annual or millenial time scales, they
> are constrained by measurements of concentrations and isotope ratios.
How a physical quantity can be "constrained by measurements"?
The quantity exists regardless of any measurements, and cannot be
constrained in that way.
> The water cycle and the carbon cycle are analogous in that the
> horizontal fluxes between geographic regions are likely to be large
> and difficult to estimate, but these must globally sum to exactly zero
> always.The system is tightly coupled in ways that are not expressed in
> regional analyses.
Same concern still stands: they must sum to zero by some
assumption of global equilibrium. But this assumption must be
proven by measurements with adequate precision and accuracy,
right? Otherwise the reasoning is circular.
> This is not a complete answer to how well constrained the main
> reservoirs actually are, but it does answer your "how is it possible"
> approach. Sometimes the best way at the whole is not to sum the parts,
If measurements in different parts were not summed in order to
obtain the global estimation, then how this number (2Gt/y) was
obtained in the first place? My undertanding is that Takahashi
has constructed a grid map with regional fluxes, see p.11
http://www.ldeo.columbia.edu/ldeo/dir/rep/PDF/bienniel_report.pdf
and then the sum of all boxes yields the difference between
global "ingassing" and "outgasing". Correct? So, again,
"how it is possible"?
- Alexi
Coby Beck
> Michael Tobis wrote:
> > On 6/1/06, Alexi Tekhasski <apredt...@austin.rr.com> wrote:
> >
> > > Now, could you tell me how it is possible to esimate global CO2 uptakes
> > > and inhales at about 100Gt/y with the above product of uncertainties,
> > > yet to arrive at a small difference (ocean uptake) of 2GT/y with an
> > > alledged error margin of only +-0.8Gt/y? Keep in mind that +-0.8Gt/y
> > > margins require estimations of two global fluxes with 0.4% accuracy,
> > > while piston velocities have 80% error margins alone. What did I
> > > miss?
> >
> > It is often the case that global quantities are better constrained
> > than regional ones.
> >
> > Consider the case of net precipitation/evaporation. Globally, it must
> > be very near zero on appropriate time scales (say, annual); in any
> > subcontinental region the uncertainty may be large.
>
> Theoretically, it must be near zero, ok. However, what guarantees
> the measurement of it to be near zero if you have only a few
> hundred measuring points, and each one has the same
> bad inaccuracy of 80%?
I don't recall what little formal training I had in scientific
measurements and significant digits etc, but isn't it the case that the
more you measure with a given uncertainty, the more accurately you know
your quantity?
> > A similar situation applies to carbon fluxes. They are globally much
> > better constrained than they are locally. Although, alas, not
> > constrained to be near zero on annual or millenial time scales, they
> > are constrained by measurements of concentrations and isotope ratios.
>
> How a physical quantity can be "constrained by measurements"?
> The quantity exists regardless of any measurements, and cannot be
> constrained in that way.
Alexi, this kind of quibble does not further a discussion. Clearly
Michael means that the range of *possible values* is "constrained by
measurements", not that we have some kind of Schrodinger's Carbon Flux
going on.
> > The water cycle and the carbon cycle are analogous in that the
> > horizontal fluxes between geographic regions are likely to be large
> > and difficult to estimate, but these must globally sum to exactly zero
> > always.The system is tightly coupled in ways that are not expressed in
> > regional analyses.
>
> Same concern still stands: they must sum to zero by some
> assumption of global equilibrium. But this assumption must be
> proven by measurements with adequate precision and accuracy,
> right? Otherwise the reasoning is circular.
No, the "assumption" was not assumed. Empirical observations show that
CO2 was in [temporaly local] equilibrium for the thousands of years
leading up to the industrial revolution. Now, empirical observations
indicate the proportion of anthro vs natural CO2 in the rising
concentration. For H20 we have sea level, ice mass, surface and ground
water measurements plus the knowledge about the relation ship between
air temperature and humidity.
Coby
Michael Tobis
from Alexi's previous challenges on sci.environment.
In any case both Coby and Alexi seem to miss my point.
Coby Beck wrote:
> Alexi Tekhasski wrote:
> > Michael Tobis wrote:
> > > The water cycle and the carbon cycle are analogous in that the
> > > horizontal fluxes between geographic regions are likely to be large
> > > and difficult to estimate, but these must globally sum to exactly zero
> > > always.The system is tightly coupled in ways that are not expressed in
> > > regional analyses.
> >
> > Same concern still stands: they must sum to zero by some
> > assumption of global equilibrium.
This is incorrect. No such assumption is required.
> > But this assumption must be
> > proven by measurements with adequate precision and accuracy,
> > right? Otherwise the reasoning is circular.
>
> No, the "assumption" was not assumed.
It's not an assumption, it's a tautology.
> Empirical observations show that
> CO2 was in [temporaly local] equilibrium for the thousands of years
> leading up to the industrial revolution. Now, empirical observations
> indicate the proportion of anthro vs natural CO2 in the rising
> concentration. For H20 we have sea level, ice mass, surface and ground
> water measurements plus the knowledge about the relation ship between
> air temperature and humidity.
This is all true but this has little to do with what I was trying to
express. I was making a far less contingent point.
The point is that what flows horizontally out of one region flows
horizontally into another, so that in the global integral these terms
are **exactly zero** under *all conditions** at **all time scales**.
This is why global measures in closed systems may have less noise than
the sum of global ones; a very large uncertainty goes away.
I did not say anythiing specifically about the terms and uncertainties
in the graph. I do think this is worth all of us looking further into,
hopefully in an honest spirit of inquiry. Perhaps we can attract some
input form carbon cycle experts.
What I said, or at least what I meant to say, was that errors in a
system with clear physical boundaries do not accumulate from errors in
a decomposition of that system into less physically delineated
components.
also...
> > How a physical quantity can be "constrained by measurements"?
> > The quantity exists regardless of any measurements, and cannot be
> > constrained in that way.
>
> Alexi, this kind of quibble does not further a discussion. Clearly
> Michael means that the range of *possible values* is "constrained by
> measurements", not that we have some kind of Schrodinger's Carbon Flux
> going on.
Indeed. I would have thought my intent was clear enough.
mt
Alexi Tekhasski
> This may be a fruitful topic, though I am surprised at how far it is
> from Alexi's previous challenges on sci.environment.
Yes, the challenge of scale separation in climate modeling
stands stronger than ever. :-)
>
> In any case both Coby and Alexi seem to miss my point.
>
[snip]
>
> This is all true but this has little to do with what I was trying to
> express. I was making a far less contingent point.
>
> The point is that what flows horizontally out of one region flows
> horizontally into another, so that in the global integral these terms
> are **exactly zero** under *all conditions** at **all time scales**.
The above statement is shokingly incorrect. If true in
reality, then there would be no global increase or
decline in CO2 concentration ever, be it glaciations
or de-glaciations.
> This is why global measures in closed systems may have less noise than
> the sum of global ones; a very large uncertainty goes away.
Do you mean "measures" in theoretical-mathematical sense,
or you equate them with "measurements"? Theoretical
measures of a conserved quantity in a closed system should
total to zero, it is by definition. However, the atmosphere
with CO2 is not a closed system, it has large reservoirs in a
form of oceans, and has sources and sinks along its surface,
sinks and sources along the land, and other sources like
anthropogenic emissions. Obviously, the sum of those sources
and sinks is not zero, othervise there would be no question
and whinning about AGW. To understand the trend of the sum,
one needs to take actual measurements of individual
components. Each individual component is a difference
between two large fluxes, the net could be poisitive, or
could be negative, you don't know apriori, your conservation
"theorem" does not apply. Therefore, you have to measure
the flux field accurately enough to determine the actual net
contribution of the component. To measure a field of the
scale of our planet, you have to "decompose" it into
individual spatio-temporal samples, measure them
individually, and sum the results up. There is no other way,
you can't cut oceans out of Earth, freeze up the corresponding
fraction of atmosphere, then evaporate only CO2
and collect it into a flask for total measurements.
> I did not say anythiing specifically about the terms and uncertainties
> in the graph. I do think this is worth all of us looking further into,
> hopefully in an honest spirit of inquiry. Perhaps we can attract some
> input form carbon cycle experts.
It would be nice to hear from them.
> What I said, or at least what I meant to say, was that errors in a
> system with clear physical boundaries do not accumulate from errors in
> a decomposition of that system into less physically delineated
> components.
The difference between theory and reality
usually is much bigger in reality than in theory.
In reality, errors do accumulate, and may do
so at a shocking rate, especially if the quantity
in question is a product of two or three
different fluctuating fields, and each field needs
to be measured in separate measurements.
Abstract physical quantities within a system
are completely different categories than
measurements of those quantities.
- aap
Raymond Arritt
Alexi Tekhasski wrote:
> Michael Tobis wrote:
> > horizontally into another, so that in the global integral these terms
> > are **exactly zero** under *all conditions** at **all time scales**.
>
> The above statement is shokingly incorrect. If true in
> reality, then there would be no global increase or
> decline in CO2 concentration ever, be it glaciations
> or de-glaciations.
You're apparently well-trained in mathematics, so it's surprising you
persist in such a fundamental error. Have you not previously
encountered Gauss' Theorem?
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > the measurement of it to be near zero if you have only a few
> > hundred measuring points, and each one has the same
> > bad inaccuracy of 80%?
>
> I don't recall what little formal training I had in scientific
> measurements and significant digits etc, but isn't it the case that the
> more you measure with a given uncertainty, the more accurately you know
> your quantity?
No, it depends. It depends on whether the quantity
is measured directly, or is a derivative from
measurements of several other basic quantities,
on spectrum of variability of those quantities, on
correlations between those measurements, and
on noise statstics (including systematic errors).
> > Same concern still stands: they must sum to zero by some
> > assumption of global equilibrium. But this assumption must be
> > proven by measurements with adequate precision and accuracy,
> > right? Otherwise the reasoning is circular.
>
> No, the "assumption" was not assumed. Empirical observations show that
> CO2 was in [temporaly local] equilibrium for the thousands of years
> leading up to the industrial revolution.
No. The same empirical observations also show that
CO2 was swinging up and down without any industrial
influence, so the alleged "equilibrium" is likely
just a "temporary local" extremum of a perpetually
oscillating function.
> Now, empirical observations
> indicate the proportion of anthro vs natural CO2 in the rising
> concentration.
Since the natural fraction of the rise in CO2 concentration
has not been quantified yet, you have no grounds to claim
any proportion or disproportion.
> For H20 we have sea level, ice mass, surface and ground
> water measurements plus the knowledge about the relation ship between
> air temperature and humidity.
I don't recall saying anything about measurements of H2O fluxes.
- aap
Michael Tobis
to be small compared to the uncertainty in the components of the sum.
As Raymond points out, my answer is analogous to the derivation of
Gauss's theorem (a.k.a. the divergence theorem).
For those readers for whom that is not helpful, let me offer an
analogy.
The number of cars in the United States is equal to the sum of the
number of cars in each of the states (and other non-overlapping
jurisdictions like DC; for clarity I will not mention those again.)
That total number changes because of the following terms: cars
constructed, cars destroyed, cars imported from other countries, cars
exported to other countries, cars crossing the border into the country,
and cars crossing the border out of the country. Each of these
quantities is easily measured.
The number of cars in a particular component of the sum, say, Illinois,
changes by those terms, plus the number of cars entering the state from
adjacent states, minus the amount of cars leaving the state into
adjacent states. These extra terms are very hard to measure. There are
numerous local streets crossing into Indiana and Wisconsin on a daily
basis, and many people commute into and out of the state (probably with
a net influx in the mornings). Also, our scenery being notoriously
dull, many people spend weekends out of the state, and our winters
being severe, many retirees spend the winters further south. There are
no borders and many exits. Therefore, the number of cars in Illinois
(not the same as the number registered here, which would spoil my
analogy) at any moment is hard to estimate.
However, every time a car moves from Illinois into Wisconsin, this has
no effect on the total number of cars in the two states. Thus,
interstate traffic, the hardest component to measure cancels out. The
estimate of the change in total number of cars in the United States is
unaffected by the main uncertainty term in the component estimate of
most of the individual states.
This doesn't mean that the number of cars in the United States doesn't
change (effectively the conclusion that Alexi seems to think I must
reach.) It means that the uncertainty of the vehicular flux, if you
will, of the nation is not the sum of the uncertainties of the
vehicular flux of the individual constituent states, and in fact may
well be smaller than the uncertainty for a specific state.
In the same way, we have global constraints on carbon fluxes that do
not apply to constituent regions. I am not as yet defending any of the
terms in the graph, except that I have confidence in the relevant
scientific community which Alexi lacks.
It is possible to conceive of a fair and civilized (as opposed to
contentious and prejudiced) examination of this matter. The question is
whether Alexi is willing to engage in it. If he is, I am willing.
However, if he fails to acknowledge elementary and straightforward
points such as this one, and instead of saying something polite like
"I'm not sure I understand you. Wouldn't that imply <absurd
conclusion>?" simply jumps to "shockingly incorrect", how exactly is a
fair-minded examination to take place?
Even if it were true that every branch of science impinging on climate
is somehow deluded or dishonest, an honest and competent skeptic ought
to be able to recognize legitimate arguments and address them, rather
than contemptuously and incorrecty dismissing them.
It has been a long time since I have seen a competent skeptic on the
public lists who plays fair. It would be nice. I learned a lot more
science from the discussions in those days. I would be very pleased if
someone were to play that role on this list, but I (and my fellow
moderators) would expect them to respect the traditional rules of
engagement in scientific controversy.
mt
EliRab...@yahoo.com
major point (besides the uncertainty of the result) is that Fan could
not differentiate between the continental US, Canada and Alaska. The
boreal forests in the later two are major sinks.
Alexi Tekhasski
> Alexi asks the question how it is possible for the uncertainty in a sum
> to be small compared to the uncertainty in the components of the sum.
> As Raymond points out, my answer is analogous to the derivation of
> Gauss's theorem (a.k.a. the divergence theorem).
>
>
> It is possible to conceive of a fair and civilized (as opposed to
> contentious and prejudiced) examination of this matter. The question is
> whether Alexi is willing to engage in it. If he is, I am willing.
>
> However, if he fails to acknowledge elementary and straightforward
> points such as this one, and instead of saying something polite like
> "I'm not sure I understand you. Wouldn't that imply <absurd
> conclusion>?" simply jumps to "shockingly incorrect", how exactly is a
> fair-minded examination to take place?
>
> Even if it were true that every branch of science impinging on climate
> is somehow deluded or dishonest, an honest and competent skeptic ought
> to be able to recognize legitimate arguments and address them, rather
> than contemptuously and incorrecty dismissing them.
>
> It has been a long time since I have seen a competent skeptic on the
> public lists who plays fair. It would be nice. I learned a lot more
> science from the discussions in those days. I would be very pleased if
> someone were to play that role on this list, but I (and my fellow
> moderators) would expect them to respect the traditional rules of
> engagement in scientific controversy.
The example about counting cars is a valid example. However,
your logical construction assumes that someone has one "big
meter" to measure the global quantity (be it change in number of
cars, or change in global amount of CO2 in atmosphere). Then
it would be true that local fluxes will cancel each other, and
there might be some accurate residual. This assumption is
not true in reality: we do not have that global meter, we cannot
strip the atmosphere along sea surface, extract all CO2, and
weight it. The CO2 fluxes were measured locally, indirectly,
and at different times, therefore the measurement errors must
add up.
More, the components of CO2 measurements are fluctuating
quantities, and, as I understand, some averaging of individual
factors was involved. This creates a sizeable potential
for a systematic error, since an average of product of fluctuating
quantities is not equal to product of their averages.
- Alexi
P.S. This is my second attempt to defend my position. The absence
of "global CO2 meter" and necessity to break measurements
down to individual fluxes makes your argument of flux canceling
irrelevant. I request to acknowledge this. You also need to admit
that your original statement of unconditional canceling of everything
on all scales is in error too, since it contradicts to known facts that
global CO2 does change, obviously due to non-zero fluxes across
atmospheric boundary, in straight accord with the Gauss Theorem.
P.P.S. Before publishing my response, I give you a permission to
edit out the whole Post Scriptum.
Thanks,
- Alexi
Michael Tobis
> The example about counting cars is a valid example.
Thank you.
> However, ...
> we do not have that global meter, we cannot
> strip the atmosphere along sea surface, extract all CO2, and
> weight it.
This is in fact incorrect, as someone with more than a passing interest
in the subject ought to be expected to know.
The atmosphere is well-mixed on the time scale of the increase;
measurements at widely scatterred sites show the same annual mean with
sufficient precision to have the net flux into the atmosphere well
constrained.
see http://cdiac.esd.ornl.gov/trends/co2/sio-keel.htm
Discussion of the uncertainties in the other terms in the decadal scale
carbon cycle is certainly of interest, but the atmospheric inventory
is, in fact, well known for these purposes.
> You also need to admit
> that your original statement of unconditional canceling of everything
> on all scales is in error too, since it contradicts to known facts ...
I will make no attempt to defend or withdraw your misunderstanding of
what I said, since I didn't say it.
The car analogy is quite precise, and makes no claims that the number
of cars is constant. It claims that the crossing of interior
boundaries, which may be a large term locally, is irrelevant in the
integral because it cancels out perfectly.
mt
Alexi Tekhasski
> Alexi Tekhasski wrote:
>
> > The example about counting cars is a valid example.
>
> Thank you.
>
> > However, ...
> > we do not have that global meter, we cannot
> > strip the atmosphere along sea surface, extract all CO2, and
> > weight it.
>
> This is in fact incorrect, as someone with more than a passing interest
> in the subject ought to be expected to know.
>
> The atmosphere is well-mixed on the time scale of the increase;
> measurements at widely scatterred sites show the same annual mean with
> sufficient precision to have the net flux into the atmosphere well
> constrained.
First, I believe that all participants (including moderators) should
obey
the imposed rules of communication and avoid personal attacks. Agree?
Second, the fact of atmosphere being globally well-mixed does
not refute nor support my statement, since the issue at hand is
about estimations of _partial_ fluxes, from oceans, and from land
(which was the original topic). The fact that about 2Gigatons of
carbon is missing from total estimated flux budget is exactly a
manifestation of problems with the current state of CO2
flux measurements.
>
> see http://cdiac.esd.ornl.gov/trends/co2/sio-keel.htm
>
> Discussion of the uncertainties in the other terms in the decadal scale
> carbon cycle is certainly of interest, but the atmospheric inventory
> is, in fact, well known for these purposes.
Again, the issue is not about global inventory, but about the rate of
its change, and interaction of sources (boundary fluxes) behind
this change. My assertion is that with given accuracy of methods
for measuring CO2 flux across global ocean, it is impossible to
estimate even the sign of the oceanic contribution.
>
> > You also need to admit
> > that your original statement of unconditional canceling of everything
> > on all scales is in error too, since it contradicts to known facts ...
>
> I will make no attempt to defend or withdraw your misunderstanding of
> what I said, since I didn't say it.
>
I really have a deep misunderstanding of what you say. You said:
"The water cycle and the carbon cycle are analogous in that the
horizontal fluxes between geographic regions are likely to be large
and difficult to estimate, but these must globally sum to exactly zero
always."
In subsequent clarification, you made even a stronger statement:
"The point is that what flows horizontally out of one region flows
horizontally into another, so that in the global integral these terms
are **exactly zero** under *all conditions** at **all time scales**.
This is why global measures in closed systems may have less noise than
the sum of global ones; a very large uncertainty goes away."
Then, when examining the analogy with cars, you said:
"This doesn't mean that the number of cars in the United States doesn't
change (effectively the conclusion that Alexi seems to think I must
reach.) "
I am really puzzled how one can reach a formal conclusion that
"the number of cars" may change if "in the global integral these terms
are **exactly zero** under *all conditions** at **all time scales**."
> The car analogy is quite precise, and makes no claims that the number
> of cars is constant. It claims that the crossing of interior
> boundaries, which may be a large term locally, is irrelevant in the
> integral because it cancels out perfectly.
First, the fluxes do not "cancel out perfectly", otherwise there will
be no observed growth in average concentration of cars [or CO2].
Second, the rate of growth is the main issue. Therefore, it is the
most relevant question how accurately can you measure the difference
between large regional fluxes, because, as I said, you have no
other means to measure them but only part-by-part.
The car analogy in not precise since it does not have some global
circulation of cars, they do not "flow horizontally". To improve
your analogy, let say that the Detroit area is a major source of
cars in the U.S., and Texas is where cars end up on junk yards,
or the sink. Then the scrap metal goes back to Detroit, and the cycle
repeats. Please keep in mind that there are other sources and
sinks in the car industry, but I want to know what is the
contribution of the US land to this picture. The question is how
accurate one can identify the difference between two large
fluxes, given that the measurements can be done only separately,
one in Detroit, one in Texas. My answer is that if the regional
measurements are inaccurate, the total will be also inaccurate
regardless of any theoretical considerations that overall everything
must almost cancel out. The question is how much is this "almost".
-Alexi
Michael Tobis
what I have said before. I made it as clear as I could.
I advise you to reread what I wrote until you understand it. My most
generous interpretation is that you may be misunderstanding the
meaning of the word "term".
mt
Coby Beck
> Michael Tobis wrote:
> > Alexi Tekhasski wrote:
> >
> > > The example about counting cars is a valid example.
> >
> > Thank you.
> >
> > > However, ...
> > > we do not have that global meter, we cannot
> > > strip the atmosphere along sea surface, extract all CO2, and
> > > weight it.
> > The atmosphere is well-mixed on the time scale of the increase;
> > measurements at widely scatterred sites show the same annual mean with
> > sufficient precision to have the net flux into the atmosphere well
> > constrained.
> Second, the fact of atmosphere being globally well-mixed does
> not refute nor support my statement, since the issue at hand is
> about estimations of _partial_ fluxes, from oceans, and from land
> (which was the original topic).
The issue, Alexi, is whether the uncertainty of the total depends on
the uncertainty of the parts. Anyway, you said there is no global
meter which clearly means, in the context of the discussion and the
analogy, that we don't know what the total flux into and out of the
atmosphere is. Now the fact that we know at any given time what the
concentration of CO2 in the atmosphere is, means there is a global
meter. This is dependent on the fact that CO2 is well mixed. There is
no argument about that.
Therefore the sum of land fluxes and oceanic fluxes is known better
than either of those parts or sub-parts.
> The fact that about 2Gigatons of
> carbon is missing from total estimated flux budget is exactly a
> manifestation of problems with the current state of CO2
> flux measurements.
Here you are making your point using knowledge that you deny having.
Without this "global meter" how do you know that 2 Gt are missing?
> > see http://cdiac.esd.ornl.gov/trends/co2/sio-keel.htm
> >
> > Discussion of the uncertainties in the other terms in the decadal scale
> > carbon cycle is certainly of interest, but the atmospheric inventory
> > is, in fact, well known for these purposes.
>
> Again, the issue is not about global inventory, but about the rate of
> its change, and interaction of sources (boundary fluxes) behind
> this change. My assertion is that with given accuracy of methods
> for measuring CO2 flux across global ocean, it is impossible to
> estimate even the sign of the oceanic contribution.
Just to take a different tack on this, the sign of the flux as well as
its magnitude can be investigated by tracking carbonic acid in the
water too.
> > > You also need to admit
> > > that your original statement of unconditional canceling of everything
> > > on all scales is in error too, since it contradicts to known facts ...
> >
> > I will make no attempt to defend or withdraw your misunderstanding of
> > what I said, since I didn't say it.
> >
>
> I really have a deep misunderstanding of what you say. You said:
>
> "The water cycle and the carbon cycle are analogous in that the
> horizontal fluxes between geographic regions are likely to be large
> and difficult to estimate, but these must globally sum to exactly zero
> always."
>
> In subsequent clarification, you made even a stronger statement:
>
> "The point is that what flows horizontally out of one region flows
> horizontally into another, so that in the global integral these terms
> are **exactly zero** under *all conditions** at **all time scales**.
> This is why global measures in closed systems may have less noise than
> the sum of global ones; a very large uncertainty goes away."
>
> Then, when examining the analogy with cars, you said:
> "This doesn't mean that the number of cars in the United States doesn't
> change (effectively the conclusion that Alexi seems to think I must
> reach.) "
>
> I am really puzzled how one can reach a formal conclusion that
> "the number of cars" may change if "in the global integral these terms
> are **exactly zero** under *all conditions** at **all time scales**."
It is just a question of figuring out when the topic is the whole
system and when it is a sub-system
> > The car analogy is quite precise, and makes no claims that the number
> > of cars is constant. It claims that the crossing of interior
> > boundaries, which may be a large term locally, is irrelevant in the
> > integral because it cancels out perfectly.
>
> First, the fluxes do not "cancel out perfectly", otherwise there will
> be no observed growth in average concentration of cars [or CO2].
You might want to consider that there is a natural system and an
anthropogenic perturbance. Before the anthropogenic disturbance there
was no significant growth or decline for many thousands of years.
> Second, the rate of growth is the main issue. Therefore, it is the
> most relevant question how accurately can you measure the difference
> between large regional fluxes, because, as I said, you have no
> other means to measure them but only part-by-part.
>
> The car analogy in not precise since it does not have some global
> circulation of cars, they do not "flow horizontally". To improve
> your analogy
[snip]
You are not improving it you are altering it so it makes your point.
Analogies are best when kept simple and targeted on illustrating an
applicable principal. Here cars=CO2, US=atmosphere, states=regional
sources and sinks. No cars are created or destroyed in the analogy
which is fine for timescales of CO2 where volcanic outgassing and
formation of limestone are not significant.
In the analogy as in the atmosphere, the states/regional fluxes are
more uncertain than the total cars/CO2
Coby
Alexi Tekhasski
> At this point you are repeating yourself. I have nothing to add to
> what I have said before. I made it as clear as I could.
Yes, it was quite clear that we have seen two contradictory
statements, so it is quite understandable that you have
nothing to add. There is a simple solution - admit your
mistake.
> I advise you to reread what I wrote until you understand it. My most
> generous interpretation is that you may be misunderstanding the
> meaning of the word "term".
>
:-) I have read all of it. It does not make sense.
I will try to approach the issue from another, simplified angle.
We have a subsystem of oceans interacting with atmosphere.
Threre are warm spots mostly emitting CO2, and there are
cold areas where CO2 sinks into oceanic conveyor.
Question for you: can you calculate from purely theoretical
grounds if the global ocean is a source or sink of CO2, and the
value of it? No. You can speculate that the two large fluxes
are likely to cancel out, but not completely, othervise there
would be no ground for debates about triggered feedbacks
etc.
Therefore, you have to measure the reality. Suppose you
have only two distinct regions: warm source, and cold sink.
Each flux is about 100Gt/y, but the measuring methodology
is known to have an error of +-80%.
Question: what is the error margin for the sum of
these fluxes, one being positive, another negative? Both
fluxes have been measured independently, with the same
equipment, but at different time? This is a simple question.
To help you, here is a relevant statement:
"Direct estimates of the net exchange rates of CO2
across the air-sea interface are associated with large
uncertainties caused by the high spatial and temporal
variability of the air-sea partial pressure difference and
uncertainties in the parameterization of air-sea gas
exchange."
This answers my "question" about "how it is possible".
The correct answer is: it is impossible, with current
measuring technique.
Cheers,
- Alexi
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > not refute nor support my statement, since the issue at hand is
> > about estimations of _partial_ fluxes, from oceans, and from land
> > (which was the original topic).
>
> The issue, Alexi, is whether the uncertainty of the total depends on
> the uncertainty of the parts. Anyway, you said there is no global
> meter which clearly means, in the context of the discussion and the
> analogy, that we don't know what the total flux into and out of the
> atmosphere is. Now the fact that we know at any given time what the
> concentration of CO2 in the atmosphere is, means there is a global
> meter. This is dependent on the fact that CO2 is well mixed. There is
> no argument about that.
>
> Therefore the sum of land fluxes and oceanic fluxes is known better
> than either of those parts or sub-parts.
If you followed the "discussion", I tried to be accurate,
my statements were related to OCEAN term, not to
the sum of everything. For example, your question was
about the LAND only. There is no global landmeter, nor
global oceanmeter. The whole issue is who is contributing,
how much, and in which direction, in order to
understand the real level of the problem.
Regarding the original statement, I misspoke.
My statement should read:
"the issue at hand is about estimations of _partial_ fluxes,
from oceans, _OR_ from land (which was the original topic)."
> Here you are making your point using knowledge that you deny having.
> Without this "global meter" how do you know that 2 Gt are missing?
I am just illustrating that the ends do not fit within current
methodology and corresponding uncertainty.
> Just to take a different tack on this, the sign of the flux as well as
> its magnitude can be investigated by tracking carbonic acid in the
> water too.
There could be many more different methods. Do you
have any results from carbonic acid method? How
accurate the measurements are?
> > "The water cycle and the carbon cycle are analogous in that the
> > horizontal fluxes between geographic regions are likely to be large
> > and difficult to estimate, but these must globally sum to exactly zero
> > always." - M.Tobis
>
> It is just a question of figuring out when the topic is the whole
> system and when it is a sub-system
There is no room for guessing. The statement was
about global carbon in the atmosphere, for the whole
system. Since the CO2 is on the rise, the sum of
fluxes over the surface of atmosphere must be positive,
in shocking accord with Gauss' Divergence Theorem.
> You might want to consider that there is a natural system and an
> anthropogenic perturbance. Before the anthropogenic disturbance there
> was no significant growth or decline for many thousands of years.
Yet many more thousand of years before there were
deglaciations and glaciations. Are you questioning
the ice core data, or what?
> > The car analogy in not precise since it does not have some global
> > circulation of cars, they do not "flow horizontally". To improve
> > your analogy
> [snip]
> You are not improving it you are altering it so it makes your point.
> Analogies are best when kept simple and targeted on illustrating an
> applicable principal. Here cars=CO2, US=atmosphere, states=regional
> sources and sinks. No cars are created or destroyed in the analogy
> which is fine for timescales of CO2 where volcanic outgassing and
> formation of limestone are not significant.
If the analogy does not hold one of the most prominent
features of the original system, it is a bad, incomplete
analogy, which serves only the point of M.Tobis. In reality,
CO2 sinks into oceans in some areas, and ocean outgases
CO2 in other areas. Therefore, cars must be considered as
having a source in Detroit, and a sink in Texas, with a scrap
metal railroad as an analogy to ocean conveyor.
> In the analogy as in the atmosphere, the states/regional fluxes are
> more uncertain than the total cars/CO2
Sum of two independently and uncertainly _measured_
quantities cannot be more certain than each
measurement.
- alexi
"Direct estimates of the net exchange rates of CO2
across the air-sea interface are associated with
large uncertainties caused by the high spatial and
temporal variability of the air-sea partial pressure
difference and uncertainties in the parameterization
of air-sea gas exchange" - Nicolas Gruber
Coby Beck
> Coby Beck wrote:
> > Alexi Tekhasski wrote:
> If you followed the "discussion", I tried to be accurate,
> my statements were related to OCEAN term, not to
> the sum of everything. For example, your question was
> about the LAND only. There is no global landmeter, nor
> global oceanmeter. The whole issue is who is contributing,
> how much, and in which direction, in order to
> understand the real level of the problem.
Yes, I have reviewed message 8 where you entered the thread. Your
question was quite specific about how is it possible that a global term
can have less uncertainty than its regional parts. For the answer, I
direct you to message 11 in the thread from Michael Tobis, it is clear
and complete.
You may well have other legitimate questions about how the global ocean
and land terms are known but you are not asking them, instead you are
resisting a very simple reality (see message 11 followups). This is
not a constructive way to converse and it results in redundancy of
content and circular converstations, to whit:
> Sum of two independently and uncertainly _measured_
> quantities cannot be more certain than each
> measurement.
Full circle. You are still wrong, although you are trying to restate
your point with artificial constraints that do not exist in reality.
So, absent any connection to the actual conversation the point as you
state it above may be reasonable. But then there is the context.
There are other ways of assessing what the sum is independently of
adding up the partial quantities.
It is all already in the thread and it is likely that the discussion
will end here unless there is something new to explore.
Coby
Eli Rabett
Alexi Tekhasski wrote:
> Coby Beck wrote:
> > Alexi Tekhasski wrote:
> > > Second, the fact of atmosphere being globally well-mixed does
> > > not refute nor support my statement, since the issue at hand is
> > > about estimations of _partial_ fluxes, from oceans, and from land
> > > (which was the original topic).
> >
> > The issue, Alexi, is whether the uncertainty of the total depends on
> > the uncertainty of the parts. Anyway, you said there is no global
> > meter which clearly means, in the context of the discussion and the
> > analogy, that we don't know what the total flux into and out of the
> > atmosphere is. Now the fact that we know at any given time what the
> > concentration of CO2 in the atmosphere is, means there is a global
> > meter. This is dependent on the fact that CO2 is well mixed. There is
> > no argument about that.
> >
> > Therefore the sum of land fluxes and oceanic fluxes is known better
> > than either of those parts or sub-parts.
>
> My statement should read:
>
> "the issue at hand is about estimations of _partial_ fluxes,
> from oceans, _OR_ from land (which was the original topic)."
>
> > Here you are making your point using knowledge that you deny having.
> > Without this "global meter" how do you know that 2 Gt are missing?
>
> I am just illustrating that the ends do not fit within current
> methodology and corresponding uncertainty.
>
its associated uncertainties, this is a bit difficult to swallow.
> > Just to take a different tack on this, the sign of the flux as well as
> > its magnitude can be investigated by tracking carbonic acid in the
> > water too.
>
> There could be many more different methods. Do you
> have any results from carbonic acid method? How
> accurate the measurements are?
The most obvious is to measure the average pH of the oceans. For any
individual sample the measurement is accurate There are variations
with depth because mixing into the deep ocean is slow (mostly it occurs
by the fall of shells from dead sea creatures). However the average pH
of the upper oceans is definitively falling, showing that more CO2 is
mixing in from the atmosphere. There are also any number of analytical
methods to measure absorbed gases and carbonate concentrations
accurately.
>
> > > "The water cycle and the carbon cycle are analogous in that the
> > > horizontal fluxes between geographic regions are likely to be large
> > > and difficult to estimate, but these must globally sum to exactly zero
> > > always." - M.Tobis
> >
> > It is just a question of figuring out when the topic is the whole
> > system and when it is a sub-system
>
> There is no room for guessing. The statement was
> about global carbon in the atmosphere, for the whole
> system. Since the CO2 is on the rise, the sum of
> fluxes over the surface of atmosphere must be positive,
> in shocking accord with Gauss' Divergence Theorem.
The pH of the oceans is decreasing as necessary if the CO2 mixing ratio
is increasing.
http://www.rsc.org/chemistryworld/News/2005/July/01070501.asp.
http://www.realclimate.org/index.php?p=169
>
> > You might want to consider that there is a natural system and an
> > anthropogenic perturbance. Before the anthropogenic disturbance there
> > was no significant growth or decline for many thousands of years.
>
> Yet many more thousand of years before there were
> deglaciations and glaciations. Are you questioning
> the ice core data, or what?
>
Those events took thousands of years. The current man made changes are
taking tens of years. Those events were driven by orbital changes.
There are no such changes taking place at this time. A LOT of fossil
fuel is being burnt, enough to account for the measured changes in CO2
concentrations in both the atmosphere and the upper oceans.
> > > The car analogy in not precise since it does not have some global
> > > circulation of cars, they do not "flow horizontally". To improve
> > > your analogy
> > [snip]
> > You are not improving it you are altering it so it makes your point.
> > Analogies are best when kept simple and targeted on illustrating an
> > applicable principal. Here cars=CO2, US=atmosphere, states=regional
> > sources and sinks. No cars are created or destroyed in the analogy
> > which is fine for timescales of CO2 where volcanic outgassing and
> > formation of limestone are not significant.
>
> If the analogy does not hold one of the most prominent
> features of the original system, it is a bad, incomplete
> analogy, which serves only the point of M.Tobis. In reality,
> CO2 sinks into oceans in some areas, and ocean outgases
> CO2 in other areas. Therefore, cars must be considered as
> having a source in Detroit, and a sink in Texas, with a scrap
> metal railroad as an analogy to ocean conveyor.
This is not even wrong. In the trivially stupid refutation, the auto
plants have moved out of Detroit and there are several in Texas.
People in Texas also buy autos. We are interested in the net. By
tracking new registrations and deregistrations we get a good measure of
the net change without actually counting each auto in the state or
country.
>
> > In the analogy as in the atmosphere, the states/regional fluxes are
> > more uncertain than the total cars/CO2
>
> Sum of two independently and uncertainly _measured_
> quantities cannot be more certain than each
> measurement.
Trivially wrong on every level. Take 100 random numbers between 0 and
1. Each is independent of the next. The uncertainty in each is of the
order of unity. The sum is are close to 50. If you do this a number
of time you find a variance that is a few percent.
Assume that a number of quantities are independently measured and are
of the same magnitude. Postulating a "true" value for each quantity,
the signs of the differences between the measured and the true values
will vary randomly, and therefore errors will cancel to a significant
extent. If each quantity is itself the result of an average of
several measurements on the same distribution you can make even
stronger statements. Same goes if you have some knowledge of the
nature of the noise on measurement of the noise for each quantity. The
sum of the errors is an outer limit and the more quantities that you
are summing, the worse the assumption.
Eli Rabett
Eli Rabett
Alexi Tekhasski wrote:
> Coby Beck wrote:
> > Alexi Tekhasski wrote:
> > > Theoretically, it must be near zero, ok. However, what guarantees
> > > the measurement of it to be near zero if you have only a few
> > > hundred measuring points, and each one has the same
> > > bad inaccuracy of 80%?
> >
> > I don't recall what little formal training I had in scientific
> > measurements and significant digits etc, but isn't it the case that the
> > more you measure with a given uncertainty, the more accurately you know
> > your quantity?
>
> No, it depends. It depends on whether the quantity
> is measured directly, or is a derivative from
> measurements of several other basic quantities,
> on spectrum of variability of those quantities, on
> correlations between those measurements, and
> on noise statstics (including systematic errors).
In all of which cases you have smaller uncertainty in the sum of the
measurements than the sum of the uncertainties in each.
>
> > > Same concern still stands: they must sum to zero by some
> > > assumption of global equilibrium. But this assumption must be
> > > proven by measurements with adequate precision and accuracy,
> > > right? Otherwise the reasoning is circular.
> >
> > No, the "assumption" was not assumed. Empirical observations show that
> > CO2 was in [temporaly local] equilibrium for the thousands of years
> > leading up to the industrial revolution.
>
> No. The same empirical observations also show that
> CO2 was swinging up and down without any industrial
> influence, so the alleged "equilibrium" is likely
> just a "temporary local" extremum of a perpetually
> oscillating function.
Note the word THOUSANDS (I might have preferred the word hundreds, but
so what).. The variations you are talking about took place on scales
of TENS OF THOUSANDS or HUNDREDS OF THOUSANDS of years. The variation
in CO2 mixing ratio since 1800 is about 100 ppm. The variation in the
10,000 years previous was about 20 ppm
http://www.grida.no/climate/ipcc_tar/wg1/fig3-2.htm. The CO2 mixing
ratio is now about 100 ppm higher than it has been in more than
400KYears at a minimum and probably quite a few million years. The
previous changes were constrained by the rates of orbital changing,
solar evolution and the evolution of life. The changes we are
currently observing are driven by burning of fossil fuels by people at
rates which increase the CO2 (and other greenhouse gases) mixing ratio
at orders of magnitude higher rates.
The RATE of change in the last 50-100 years is huge compared to
previous times which makes a joke of your reference to a perpetually
oscillating function.
>
> > Now, empirical observations
> > indicate the proportion of anthro vs natural CO2 in the rising
> > concentration.
>
> Since the natural fraction of the rise in CO2 concentration
> has not been quantified yet, you have no grounds to claim
> any proportion or disproportion.
>
Again, you display your ignorance. Isotope studies and other methods
fingerprint the source of the excess CO2. See for example
http://www.radix.net/~bobg/faqs/scq.CO2rise.html for a simple primer
> > For H20 we have sea level, ice mass, surface and ground
> > water measurements plus the knowledge about the relation ship between
> > air temperature and humidity.
>
> I don't recall saying anything about measurements of H2O fluxes.
>
So? Well at least then you didn't say anything wrong. Be thankful for
small mercies.
Eli Rabett
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > I am just illustrating that the ends do not fit within current
> > methodology and corresponding uncertainty.
> >
> Since you show no familiarity with what the current methodology is and
> its associated uncertainties, this is a bit difficult to swallow.
Please try to swallow an example I gave in response to last Coby's
comment.
> > There could be many more different methods. Do you
> > have any results from carbonic acid method? How
> > accurate the measurements are?
>
> The most obvious is to measure the average pH of the oceans. For any
> individual sample the measurement is accurate There are variations
> with depth because mixing into the deep ocean is slow (mostly it occurs
> by the fall of shells from dead sea creatures). However the average pH
> of the upper oceans is definitively falling, showing that more CO2 is
> mixing in from the atmosphere. There are also any number of analytical
> methods to measure absorbed gases and carbonate concentrations
> accurately.
World ocean is vast, there are spatial variations. How do you
measure the average if you have only few stations/ships?
Aha, you use continuity assumptions and grid
approximations...anything else?
> > There is no room for guessing. The statement was
> > about global carbon in the atmosphere, for the whole
> > system. Since the CO2 is on the rise, the sum of
> > fluxes over the surface of atmosphere must be positive,
> > in shocking accord with Gauss' Divergence Theorem.
>
> The pH of the oceans is decreasing as necessary if the CO2 mixing ratio
> is increasing.
> http://www.rsc.org/chemistryworld/News/2005/July/01070501.asp.
> http://www.realclimate.org/index.php?p=169
Ok, there might be indications that oceans consume CO2,
but they were doing this for the last few billion years, so what?
And how it is related to my particular statement
about global carbon increase in atmosphere?
> > Yet many more thousand of years before there were
> > deglaciations and glaciations. Are you questioning
> > the ice core data, or what?
> >
> Those events took thousands of years.
There are no evidence that there were no abrupt changes
in CO2 concentrations in the distant past since all
known records are essentially smoothened over thousand
of years: worms and turbulence stirred sediments, and
percolation within firn takes hundreds of years as well.
> The current man made changes are
> taking tens of years. Those events were driven by orbital changes.
This is a theory that have a lot of holes to explain. Large volcano
eruptions also had happened over weeks or months.
> There are no such changes taking place at this time. A LOT of fossil
> fuel is being burnt, enough to account for the measured changes in CO2
> concentrations in both the atmosphere and the upper oceans.
Yes, fossil fulels are burning. But how much is a LOT? Is the natural
level of 2000 ppm a lot? Or what about 6000 ppm?
> > If the analogy does not hold one of the most prominent
> > features of the original system, it is a bad, incomplete
> > analogy, which serves only the point of M.Tobis. In reality,
> > CO2 sinks into oceans in some areas, and ocean outgases
> > CO2 in other areas. Therefore, cars must be considered as
> > having a source in Detroit, and a sink in Texas, with a scrap
> > metal railroad as an analogy to ocean conveyor.
>
> This is not even wrong. In the trivially stupid refutation, the auto
> plants have moved out of Detroit and there are several in Texas.
> People in Texas also buy autos. We are interested in the net. By
> tracking new registrations and deregistrations we get a good measure of
> the net change without actually counting each auto in the state or
> country.
World ocean has distinctive areas where it outgases CO2, and where
it sinks CO2. Detroit and Texas are just an attempt of symbolic
representation of that fact in the car analogy. Taking a measurement
from a ship station is equivalent of a DWI cordon on a county
road once a month. Also, you don't have tax collection offices
on every 100x100 mile sea area all the time. Therefore, I wholly
agree with your characterization of your own refutation.
> > Sum of two independently and uncertainly _measured_
> > quantities cannot be more certain than each
> > measurement.
>
> Trivially wrong on every level. Take 100 random numbers between 0 and
> 1. Each is independent of the next. The uncertainty in each is of the
> order of unity. The sum is are close to 50. If you do this a number
> of time you find a variance that is a few percent.
>
> Assume that a number of quantities are independently measured and are
> of the same magnitude. Postulating a "true" value for each quantity,
> the signs of the differences between the measured and the true values
> will vary randomly, and therefore errors will cancel to a significant
> extent. If each quantity is itself the result of an average of
> several measurements on the same distribution you can make even
> stronger statements. Same goes if you have some knowledge of the
> nature of the noise on measurement of the noise for each quantity. The
> sum of the errors is an outer limit and the more quantities that you
> are summing, the worse the assumption.
Do you know the nature of noise and 80% scatter in estimation
of the stagnant film coefficient? When you will learn it, I am all
ears...
Also, measurements in ocean were taken essentially once per
ship station, so they hardly can qualify as tossing the same coin
100 times. More, the quantities in ocean do vary substantially,
therefore your assumption "of the same magnitude" is incorrect.
You also are not taking into account systematic errors that
result from correlations between wind direction, velocity, and
partial CO2 pressure.
- alexi
Alexi Tekhasski
> Yes, I have reviewed message 8 where you entered the thread. Your
> question was quite specific about how is it possible that a global term
> can have less uncertainty than its regional parts. For the answer, I
> direct you to message 11 in the thread from Michael Tobis, it is clear
> and complete.
Yes, the message #11 is clear in the sense that it shows
complete misappropriation of the whole issue. My
rhetorical question was about ocean fluxes, and ocean
fluxes only. Carbon flux across sea surface IS NOT
GLOBAL, there are other areas with equally unknown fluxes,
therfore the whole Tobis' construction **does not apply**
in first place.
>
> You may well have other legitimate questions about how the global ocean
> and land terms are known but you are not asking them, instead you are
> resisting a very simple reality (see message 11 followups). This is
> not a constructive way to converse and it results in redundancy of
> content and circular converstations, to whit:
>
> > Sum of two independently and uncertainly _measured_
> > quantities cannot be more certain than each
> > measurement.
>
> Full circle. You are still wrong, although you are trying to restate
> your point with artificial constraints that do not exist in reality.
No, I am just trying to explain where your (M.Tobis) reasoning
is wrong. I even gave you a quotation from CO2 authority,
which you elected to ignore. Once again, ocean's subsystem
is only a part of global flux. Since you don't know what the
other fluxes are (your original question is the proof), you cannot
apply the "global constraint" argument.
> So, absent any connection to the actual conversation the point as you
> state it above may be reasonable.
Incorrect. My statement is not only "reasonable", it is
an undisputed fact from the general theory of
measurements.
> But then there is the context.
As I have shown above, the "context" does not apply.
> There are other ways of assessing what the sum is independently of
> adding up the partial quantities.
No, while in general it is possible, this is incorrect with regard
to our particular case of CO2 flux measurements..
The fact that global quantity is constrained
by indirect evidence does not make inaccurate measurements of
partial quantities any more accurate. You keep ignoring that
I am talking about partial flux from world oceans. Your own
original question was about US land. There are other lands,
with different average value of CO2 flux. Only the total of all
fluxes is "constrained" by evidence of average concentration.
You don't have any other quantitative idea about how large
or small the other fluxes are, other than
it is likely that the difference between positive and negative
regional fluxes must be smaller than each regional flux if
taken seprately, which is obvious.Your own question
manifests the fact that you have no theoretical idea about
how large the partial U.S. land contribution is. Therefore,
the only way is to try to measure it, and the current way
is to measure and add up partial quantities from
smaller regions. You do not have any other means to
assess the sum of ocean's fluxes other than measure
them region by region, and then sum them up.
> It is all already in the thread and it is likely that the discussion
> will end here unless there is something new to explore.
Yes, there is a lot to expolre. For example, you elected
to ignore the citation I gave, about "uncertainties caused
by the high spatial and temporal variability of the air-sea
partial pressure difference". You already confronted me
once about my statement that all measurements must
be conducted on at least half-our basis. To give you an
idea about what kind of uncertainty is at stake, let me
construct an example about CO2 flux measuremetns.
>From literature is follows that the sea-air flux is
proportional to effective difference in CO2 concentrations
across air-water interface (say, dC), and wind velocity W
( in first approximation). So, we have:
Flux = k * W * dC;
where k is a coefficient related to ill-known "stangnant
film" property.
The task is to determine an average flux over a period of one
year, <Flux>. Since measurements of W and dC are not
taken every minute, and not simultaneously, I assume that
some sort of individual averages of W and dC are used in
calculations. Now, imagine the following realistic weather
conditions over one square meter of sea surface. Let say
that the surface is somewhere in middle Atlantic, and
90 out of 100 seconds the wind does not blow (W=0), all
is calm and equilibrated, so dC is also near zero, and terefore
no Flux is coming in neither direction. But in the last 10
seconds the wind gets to 10m/s, and brings a variation
in dC, say, 10 units (whatever they are). Then the gusts
repeat whole days or weeks.
So, the average wind would be 10/10=1m/s, and average
difference in concentrations would be 10/10 = 1 unit. As result,
the estimate of average flux would be k*1*1 = k units per
square meter. This is in accord to current methodology.
Now, what happens in reality? Over 90 seconds the instant
flux is zero, but during the gust the instant flux is k*10*10 =
100k. Average of the flux (1:10) gives you a correct answer,
<Flux> = 10*k. As you can see, the real flux is 10x larger
than the initial methodology would assess. Of course, this
is an abstracted example, but you get an idea why
"high spatial and temporal variability" may cause such
alarming uncertainties if the measurements are not
conducted in accord with spatio-temporal characteristics
of that variability.
I also have serious reservation about uniformity of the
stagnant film coefficient k, but I will address it later.
- Alexi
Eli Rabett
> Eli Rabett wrote:
> > Alexi Tekhasski wrote:
> > > I am just illustrating that the ends do not fit within current
> > > methodology and corresponding uncertainty.
> > >
> > Since you show no familiarity with what the current methodology is and
> > its associated uncertainties, this is a bit difficult to swallow.
>
> Please try to swallow an example I gave in response to last Coby's
> comment.
Right, the magisterial and meaningless wave of your hand. What
specific example did you give? How do you think that it defeated
Coby's comment?
>
> > > There could be many more different methods. Do you
> > > have any results from carbonic acid method? How
> > > accurate the measurements are?
> >
> > The most obvious is to measure the average pH of the oceans. For any
> > individual sample the measurement is accurate There are variations
> > with depth because mixing into the deep ocean is slow (mostly it occurs
> > by the fall of shells from dead sea creatures). However the average pH
> > of the upper oceans is definitively falling, showing that more CO2 is
> > mixing in from the atmosphere. There are also any number of analytical
> > methods to measure absorbed gases and carbonate concentrations
> > accurately.
>
> World ocean is vast, there are spatial variations. How do you
> measure the average if you have only few stations/ships?
> Aha, you use continuity assumptions and grid
> approximations...anything else?
You study mixing in the upper ocean, ocean currents and many other
things that show that this is accurate. You take sets of measurments.
You study the variations. You understand their cause. You note the
range of the variations. A good introduction can be found at
http://www.royalsoc.ac.uk/displaypagedoc.asp?id=13539 On the other
hand, one can adopt your tactic and furiously wave your hands.
>
> > > There is no room for guessing. The statement was
> > > about global carbon in the atmosphere, for the whole
> > > system. Since the CO2 is on the rise, the sum of
> > > fluxes over the surface of atmosphere must be positive,
> > > in shocking accord with Gauss' Divergence Theorem.
> >
> > The pH of the oceans is decreasing as necessary if the CO2 mixing ratio
> > is increasing.
> > http://www.rsc.org/chemistryworld/News/2005/July/01070501.asp.
> > http://www.realclimate.org/index.php?p=169
>
> Ok, there might be indications that oceans consume CO2,
> but they were doing this for the last few billion years, so what?
You appear unaware that the pH of the oceans has been fairly constant
at about 8.2 for millions of years, that life in the oceans has evolved
under those conditions and many organisms can only function within a
narrow range of pH (ps talk to your physician and ask what the pH of
your blood is and how fast you die if it changes by a little bit).
Even though there is significant buffering the current changes in the
pH of the oceans is rapidly approaching a dangerous level for many
organisms. See the Royal Society Report that I provided the URL for.
Again, you combine arrogance with ignorance in a dangerous mixture.
> And how it is related to my particular statement
> about global carbon increase in atmosphere?
>
Because of the interactions between the atmosphere and the oceans an
increase in pH is a sure sign that atmospheric CO2 is increasing.
Changes in isotopic ratios are excellent markers of the sources.
> > > Yet many more thousand of years before there were
> > > deglaciations and glaciations. Are you questioning
> > > the ice core data, or what?
> > >
> > Those events took thousands of years.
>
> There are no evidence that there were no abrupt changes
> in CO2 concentrations in the distant past since all
> known records are essentially smoothened over thousand
> of years: worms and turbulence stirred sediments, and
> percolation within firn takes hundreds of years as well.
>
You are simply guessing. Guessing wrong, but guessing. Guess what,
you may impress your friends at the bar, but there are people in this
forum who actually do, read and think about the available data and
records.
> > The current man made changes are
> > taking tens of years. Those events were driven by orbital changes.
>
> This is a theory that have a lot of holes to explain.
That the changes in CO2 concentrations are taking tens of years? Come
now, even you cannot be so purposfully ignorant of the facts. If you
wish to examine the data, see http://tinyurl.com/mqsvm and other files
on the cdiac site.
> Large volcano eruptions also had happened over weeks or months.
>
One tends to notice large volcanos. Moreover on the scale of eruptions
that have happened in historical times, there have not been large steps
in CO2 concentrations. For example, try and find Pinatubo in the Mauna
Loa record. Maybe something on the scale of the Yellowstone eruptions
2 million years ago, but again, the paleoclimate record shows you are
waving your hands.
> > There are no such changes taking place at this time. A LOT of fossil
> > fuel is being burnt, enough to account for the measured changes in CO2
> > concentrations in both the atmosphere and the upper oceans.
>
> Yes, fossil fulels are burning. But how much is a LOT? Is the natural
> level of 2000 ppm a lot? Or what about 6000 ppm?
Are you purposely trying to put forth pathetic arguments? First of all
what do you mean by the natural level of 2000 ppm or 6000 ppm. Natural
under what conditions? When levels were that high, the earth was a
very different place and not very hospitable for thee and me
>
> > > If the analogy does not hold one of the most prominent
> > > features of the original system, it is a bad, incomplete
> > > analogy, which serves only the point of M.Tobis. In reality,
> > > CO2 sinks into oceans in some areas, and ocean outgases
> > > CO2 in other areas. Therefore, cars must be considered as
> > > having a source in Detroit, and a sink in Texas, with a scrap
> > > metal railroad as an analogy to ocean conveyor.
> >
> > This is not even wrong. In the trivially stupid refutation, the auto
> > plants have moved out of Detroit and there are several in Texas.
> > People in Texas also buy autos. We are interested in the net. By
> > tracking new registrations and deregistrations we get a good measure of
> > the net change without actually counting each auto in the state or
> > country.
>
> World ocean has distinctive areas where it outgases CO2, and where
> it sinks CO2. Detroit and Texas are just an attempt of symbolic
> representation of that fact in the car analogy. Taking a measurement
> from a ship station is equivalent of a DWI cordon on a county
> road once a month. Also, you don't have tax collection offices
> on every 100x100 mile sea area all the time. Therefore, I wholly
> agree with your characterization of your own refutation.
>
You would make a very bad traffic engineer. You do not have to measure
the traffic on every road all the time to estimate traffic in and out
of a location.
> > > Sum of two independently and uncertainly _measured_
> > > quantities cannot be more certain than each
> > > measurement.
An absolute statement not associated with any specific example and thus
falsifiable by counterexample. Here is one
> >
> > Trivially wrong on every level. Take 100 random numbers between 0 and
> > 1. Each is independent of the next. The uncertainty in each is of the
> > order of unity. The sum is are close to 50. If you do this a number
> > of time you find a variance that is a few percent
Here is another
> >
> > Assume that a number of quantities are independently measured and are
> > of the same magnitude. Postulating a "true" value for each quantity,
> > the signs of the differences between the measured and the true values
> > will vary randomly, and therefore errors will cancel to a significant
> > extent. If each quantity is itself the result of an average of
> > several measurements on the same distribution you can make even
> > stronger statements. Same goes if you have some knowledge of the
> > nature of the noise on measurement of the noise for each quantity. The
> > sum of the errors is an outer limit and the more quantities that you
> > are summing, the worse the assumption.
>
> Do you know the nature of noise and 80% scatter in estimation
> of the stagnant film coefficient? When you will learn it, I am all
> ears...
A retreat to a specific, ill defined and totally meaningless statment.
Clearly Alexi cannot show that the two (of many) counterexamples
offered above are incorrect.
Your statement which you have repeated across many posts was absolute.
My first example was of a set of random numbers. If you want to talk
about heat flow across boundary layers, feel free, but you made an
absolute statement about errors adding that is trivially falsified.
Now you are dragging the cat in.
> Also, measurements in ocean were taken essentially once per
> ship station,
Again, an irrelevancy. An incorrect irrelevancy, but what the heck.
> so they hardly can qualify as tossing the same coin
> 100 times. More, the quantities in ocean do vary substantially,
> therefore your assumption "of the same magnitude" is incorrect.
Wrong. Temperature and pH vary by less than an order of magnitude.
pH, for example varies by .3 units at most which is about a factor of
two. Much less on the scale of hundreds of km. Temperatures vary by
about 30% from mimima to maxima (I use the Kelvin scale for
temperature, which is the appropriate one, not an arbitrary scale such
as Celcius). Again, much less on the scale of hundreds of miles.
> You also are not taking into account systematic errors that
> result from correlations between wind direction, velocity, and
> partial CO2 pressure.
CO2 and other trace gases are characterized by mixing ratios by volume.
Take a look at the Mauna Loa record for example. This measure is
independent of wind induced pressure differences. Measurement stations
are located in remote areas to eliminate possible effects from trace
gas point sources. Grab a clue.
Oh yes, please don't try and drag out the Mauna Loa is a volcano line.
The effects of emissions from the volcano at the measuring station have
been studied in detail and are known to be small. They are mostly
below 0.1 ppm and only after an eruption have they been larger.
http://www.mlo.noaa.gov/HISTORY/PUBLISH/steve/VolcCO2.htm .
Besides which, there is no volcano at Barrow, and other sites which
show the same trends as at MLO
You are guessing Alexi. That won't work here, even for someone as
aggressive as you.
Eli Rabett
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > Eli Rabett wrote:
> > > Alexi Tekhasski wrote:
> > > > I am just illustrating that the ends do not fit within current
> > > > methodology and corresponding uncertainty.
> > > >
> > > Since you show no familiarity with what the current methodology is and
> > > its associated uncertainties, this is a bit difficult to swallow.
> >
> > Please try to swallow an example I gave in response to last Coby's
> > comment.
>
> Right, the magisterial and meaningless wave of your hand. What
> specific example did you give?
Read the example again, it was very specific. It shows that the
estimate of CO2 flux could be anything if you do not take into
account fine time correlations between measured quantities on
actual time scales. Depending on the shape of correlation, the
result can be anything, from a small fraction of what the product
of averages gives you, to 10x and more.
> How do you think that it defeated Coby's comment?
I don't think it defeated Coby's comment since the comment
was refuted in upper part of my post, and in few previous
posts. My new specific example illustrates NEW side of the
issue, not discussed before. It demonstrates the level of
uncertainties mentioned in the reference I already posted:
"Direct estimates of the net exchange rates of CO2
across the air-sea interface are associated with
large uncertainties caused by the high spatial and
temporal variability of the air-sea partial pressure
difference" - Samiente, Gloor, & Gruber:
http://www.uib.no/jgofs/Publications/other_pub/Southampton_Sarmiento.pdf
Cheers,
- Alexi
Michael Tobis
across the air-sea interface are associated with
large uncertainties caused by the high spatial and
temporal variability of the air-sea partial pressure
> It demonstrates the level of
> uncertainties mentioned in the reference I already posted:
The abstract makes no claim of relevance to the global carbon cycle.
Instead, it is an effort to better constrain the general circulation
of the ocean, a dramatically different objective.
It is entirely irrelevant to your original question of whether global
inventories are well-constrained.
I think we have satisfactorily argued that the atmosphere inventory is
well-constrained, and Eli has made a case that the ocean inventory is
well constrained, for which I thank him, and which I intend to study
with some care.
This being the case, as explained in tedious detail already and
grudgingly half-conceded by you, uncertainties in local fluxes,
however important for other reasons, do not affect global inventory
assessments, and globally integrated fluxes can be backed out of the
global inventroies without reference to local uncertainties.
Indeed, the quoted abstract has the following sentence:
"However, our results show a shift of uptake from the high to the
mid-latitudes, with far less uptake in the Southern Ocean"
which is consistent with an accepted constraint on globally integrated flux.
> My new specific example illustrates NEW side of the
> issue, not discussed before.
It would be very helpful if you would take the trouble to define "the
issue" as you see it, as well as whatever new "side" you choose to
bring up.
I thought we were discussing the existence of otherwise of useful
measures of the components of the contemporary carbon cycle. Is this,
in fact, what we are discussing, or not? If not, please start a new
thread with a suitable topic and context so we stop going round in
circles. You do want the conversation to make progress, right?
Keyword searching in abstracts is a known, established trick of the
obfuscators. So is nitpicking and refusal to stay on topic. So is a
sophisticated vocabulary arrayed in support of unsophisticated
theories that are inconsistent with evidence.
If you are here simply to obfuscate the conversation, I merely ask you
to check in with your conscience before proceeding.
mt
Alexi Tekhasski
> "Direct estimates of the net exchange rates of CO2
> across the air-sea interface are associated with
> large uncertainties caused by the high spatial and
> temporal variability of the air-sea partial pressure
> difference" - Samiente, [sic] Gloor, & Gruber:
>
> > It demonstrates the level of
> > uncertainties mentioned in the reference I already posted:
>
> The abstract makes no claim of relevance to the global carbon cycle.
> Instead, it is an effort to better constrain the general circulation
> of the ocean, a dramatically different objective.
>
> It is entirely irrelevant to your original question of whether global
> inventories are well-constrained.
You are sadly mistaken. It was never my question.
Please review the whole thread.
It started with Coby's confusion about the contribution
of US region to carbon cycle. Note: _regional_ contribution.
I responded with a similar problem of another _regional_
contribution, from oceans, and expressed concern about
feasibility of reliable measurements across sea-air
interface and associated error in estimation of a small
difference between several large fluxes from different
sub-regions of the world ocean.
It was you who changed the topic to "global constraint",
Knowledge of "global constraint" does not lead to understanding
of causes that lead to global changes, as I will explain
below.
> I think we have satisfactorily argued that the atmosphere inventory is
> well-constrained, and Eli has made a case that the ocean inventory is
> well constrained, for which I thank him, and which I intend to study
> with some care.
>
> This being the case, as explained in tedious detail already and
> grudgingly half-conceded by you, uncertainties in local fluxes,
> however important for other reasons, do not affect global inventory
> assessments, and globally integrated fluxes can be backed out of the
> global inventroies without reference to local uncertainties.
This is a pretty mioptic point of view. First, global "inventory"
is not inaffected by local fluxes, it is _defined_ by them, by an
integral over planetary surface and time. Second, if someone
wants to understand the _trend_ in global inventory (which I guess
is the primary purpose of the whole climatology), one has
to study local fluxes. A posteriory knowledge of changes in
only global CO2 inventory does not serve any prognostic
purpose.
>
> Indeed, the quoted abstract has the following sentence:
>
> "However, our results show a shift of uptake from the high to the
> mid-latitudes, with far less uptake in the Southern Ocean"
>
> which is consistent with an accepted constraint on globally integrated flux.
You continue to bark on a wrong tree, and continue to miss the point.
More, your conclusion from the above sentence is wrong,
since the atmospheric system over oceans is not a closed
system, and there might be _local_ exchanges of CO2 fluxes
with nearby land. Therefore, no conclusion could be drawn
about consistency or inconsistency of the observation
with "constraints on globally integrated flux".
>
> > My new specific example illustrates NEW side of the
> > issue, not discussed before.
>
> It would be very helpful if you would take the trouble to define "the
> issue" as you see it, as well as whatever new "side" you choose to
> bring up.
>
> I thought we were discussing the existence of otherwise of useful
> measures of the components of the contemporary carbon cycle. Is this,
> in fact, what we are discussing, or not? If not, please start a new
> thread with a suitable topic and context so we stop going round in
> circles. You do want the conversation to make progress, right?
I thought it was quite clear. First, I brougt to your attention that
an algebraic sum of highly-uncertain measurments taken in few
isolated places over the time span of 8 years cannot yield a good
accurate total. The second issue is about the potential inaccuracy
of those individual measurements, and overlooked fundamental reasons
behind the uncertainty. The third issue with the same problem
could be about mechanisms behind the "parametrization"
of the sea-air, interface, i.e. about physics of "stangant film"
approximation.
I probably need to spell out what I think the issue is. IMO,
the issue is about the trend in global inventory of CO2 in
atmosphere. Merely stating that we can measure the global
inventory does not answer the main question about its
dynamics. To understand the dynamics, one needs to
analyse components of the flux that drives the global
inventory. That's why we need to measure individual
fluxes, and there are problems associated with those
measurements. That's why the knowledge of "global
constraint" does not advance any understanding of
the problem. That's why big money are spent on
oceanic expeditions to conduct local measurements.
That's why your position does not make any sense
(other than obvious restatement of Gauss theorem).
> Keyword searching in abstracts is a known, established trick of the
> obfuscators. So is nitpicking and refusal to stay on topic. So is a
> sophisticated vocabulary arrayed in support of unsophisticated
> theories that are inconsistent with evidence.
>
> If you are here simply to obfuscate the conversation, I merely ask you
> to check in with your conscience before proceeding.
It is funny. I initiated this particular conversation, at least the
most
substantial part of it, and now you are accusing me of obfuscaion of
my own conversation? It does not sound as a coherent thought from you.
If you want to talk to yourself and yourself only, with self-imposed
constraints of misconceptions and rigid belief system, it is advisable
to check with your local church.
- aap
Coby Beck
> Michael Tobis wrote:
> >
> > inventories are well-constrained.
>
> You are sadly mistaken. It was never my question.
> Please review the whole thread.
> It started with Coby's confusion about the contribution
> of US region to carbon cycle. Note: _regional_ contribution.
> I responded with a similar problem of another _regional_
> contribution, from oceans, and expressed concern about
> feasibility of reliable measurements across sea-air
> interface and associated error in estimation of a small
> difference between several large fluxes from different
> sub-regions of the world ocean.
Alexi, with all due respect, it is ridiculous to consider the ocean as
a "regional" component of the carbon cycle system. It is clearly most
sensible to divide it into land, air and ocean, and then subdivide
those components by regions. If your "ocean as a region" model is in
fact how you are truly thinking about the problem, then I suggest you
start over. Personally, I think you are just playing word games. Your
original post was very clear in asking how a global flux could be more
precisely known than any of its components.
Can we please put that to rest and if you still have some *other*
issue, restate it as a new topic?
> It was you who changed the topic to "global constraint",
> Knowledge of "global constraint" does not lead to understanding
> of causes that lead to global changes, as I will explain
> below.
No, see above.
> > I think we have satisfactorily argued that the atmosphere inventory is
> > well-constrained, and Eli has made a case that the ocean inventory is
> > well constrained, for which I thank him, and which I intend to study
> > with some care.
> >
> > This being the case, as explained in tedious detail already and
> > grudgingly half-conceded by you, uncertainties in local fluxes,
> > however important for other reasons, do not affect global inventory
> > assessments, and globally integrated fluxes can be backed out of the
> > global inventroies without reference to local uncertainties.
>
> This is a pretty mioptic point of view. First, global "inventory"
> is not inaffected by local fluxes, it is _defined_ by them, by an
> integral over planetary surface and time.
Again, respectfully, this is a very annoying tactic and one that raises
a lot of red flags as to wether or not you are sincere. mt said
"uncertainties in local fluxes [...] do not affect global inventory
assessments" and he is obviously talking about measurements. You
respond with a trivially obvious "correction" as if mt was saying
something in conflict with it. That does not advance a conversation.
> Second, if someone
> wants to understand the _trend_ in global inventory (which I guess
> is the primary purpose of the whole climatology), one has
> to study local fluxes
Why do you persist in this ridiculous argument? If you want to know
the trend in global inventory of atmospheric carbon dioxide just go
look at the Mauna Loa records. Study the change in carbonic acid in
the ocean.
Stop with this red herring.
> > If you are here simply to obfuscate the conversation, I merely ask you
> > to check in with your conscience before proceeding.
>
> It is funny. I initiated this particular conversation, at least the
> most
> substantial part of it, and now you are accusing me of obfuscaion of
> my own conversation? It does not sound as a coherent thought from you.
> If you want to talk to yourself and yourself only, with self-imposed
> constraints of misconceptions and rigid belief system, it is advisable
> to check with your local church.
This is just a bit bizarre, frankly. "Your own conversation"?? And
it's gratuitously rude. Everytime some one throws this "religious" mud
it just stinks of cheap shots and a poor effort at creating an alibi
for their own unwillingness to learn.
Coby
Michael Tobis
> if someone
> wants to understand the _trend_ in global inventory
Are you suggesting that we don't understand the trend to first order?
> (which I guess
> is the primary purpose of the whole climatology),
How silly.
No, it's pretty much an established result in *geochemistry*, and
isn't really a *climatological* result at all. You may wish to gain
more understanding of fields which you are so quick to criticize.
> [to understand the _trend_ in global inventory] ...
> one has
> to study local fluxes.
I think this is trivially and obviously incorrect, for reasons which
have been explained in great detail.
To clarify what it is we are discussing, do you agree that we can be
confident that the most recent part of this signal is anthropogenic?
http://www.globalwarmingart.com/wiki/Image:Carbon_Dioxide_400kyr_Rev.png
Do you contend that no global evidence could possibly suffice for such
attribution?
mt
Alexi Tekhasski
> Alexi Tekhasski wrote:
> > Michael Tobis wrote:
> > >
> > > It is entirely irrelevant to your original question of whether global
> > > inventories are well-constrained.
> >
> > You are sadly mistaken. It was never my question.
> > Please review the whole thread.
> > It started with Coby's confusion about the contribution
> > of US region to carbon cycle. Note: _regional_ contribution.
> > I responded with a similar problem of another _regional_
> > contribution, from oceans, and expressed concern about
> > feasibility of reliable measurements across sea-air
> > interface and associated error in estimation of a small
> > difference between several large fluxes from different
> > sub-regions of the world ocean.
> Alexi, with all due respect, it is ridiculous to consider the ocean as
> a "regional" component of the carbon cycle system. It is clearly most
> sensible to divide it into land, air and ocean, and then subdivide
> those components by regions. If your "ocean as a region" model is in
> fact how you are truly thinking about the problem, then I suggest you
> start over.
Then why don't you suggest to withdraw and rewrite all textbooks
that depicted the global carbon cycle as partitioned into land and
ocean? BTW, combining different regional partition into bigger
divisions does not change the fact that the regional fluxes are
high and still are hard-to-impossible to measure directly.
> .. Personally, I think you are just playing word games. Your
> original post was very clear in asking how a global flux could be more
> precisely known than any of its components.
Since you have suspected me of playing "word games",
I begin to suspect that you are intentionally dull. In my
post I specifically expressed my concern about accuracy
of estimation of ocean uptake of 2.1Gt/y +-0.8Gt/y
Therfore, it is obvious that all my statements and
concerns were about uncertainties in sea-air exchange.
If you (or Mr.Tobis and Prof. Arritt) misconstrued my posts
and jumped into meanigless discussion of "global constraints",
it is your problem.
> Can we please put that to rest and if you still have some *other*
> issue, restate it as a new topic?
>
> > It was you who changed the topic to "global constraint",
> > Knowledge of "global constraint" does not lead to understanding
> > of causes that lead to global changes, as I will explain
> > below.
>
> No, see above.
Yes, see above.
> > > I think we have satisfactorily argued that the atmosphere inventory is
> > > well-constrained, and Eli has made a case that the ocean inventory is
> > > well constrained, for which I thank him, and which I intend to study
> > > with some care.
> > >
> > > This being the case, as explained in tedious detail already and
> > > grudgingly half-conceded by you, uncertainties in local fluxes,
> > > however important for other reasons, do not affect global inventory
> > > assessments, and globally integrated fluxes can be backed out of the
> > > global inventroies without reference to local uncertainties.
> >
> > This is a pretty myopic point of view. First, global "inventory"
> > is not inaffected by local fluxes, it is _defined_ by them, by an
> > integral over planetary surface and time.
>
> Again, respectfully, this is a very annoying tactic and one that raises
> a lot of red flags as to wether or not you are sincere. mt said
> "uncertainties in local fluxes [...] do not affect global inventory
> assessments" and he is obviously talking about measurements. You
> respond with a trivially obvious "correction" as if mt was saying
> something in conflict with it. That does not advance a conversation.
There is no tactic, and I am sincerely trying to help you.
This is my last attempt. Say, you have a global balance,
G = A + B + C. Knowing G(t) (and maybe A) does not
answer what B(t) and C(t) are, and how they can drive G in
future. Both B and C have their own dynamics, sensitivities,
and time lags with regard to temperatures and
CO2 flux. Without knowing B and C separately, it is meanigless
to speculate about trends in global increase, and capacities
of B and C to uptake CO2. More, it is not entirely impossible
that in reality the whole A is totally consumed by B, but the
whole rise in G is provided by changes in C.
Therefore, my correction is not trivial, it
emphasizes entirely different and meaningful posing
of the whole question. Yes, uncertainties in _measurements_
of local fluxes do not affect independent measurement of
global inventory, but the latter does not answer important questions.
It is just like trying to say that total mass of Earth is constant
(barring continous loss of hydrogen into open space), it answers
all other questions, and therefore we don't need to measure
anything else.
> > Second, if someone
> > wants to understand the _trend_ in global inventory (which I guess
> > is the primary purpose of the whole climatology), one has
> > to study local fluxes
>
> Why do you persist in this ridiculous argument? If you want to know
> the trend in global inventory of atmospheric carbon dioxide just go
> look at the Mauna Loa records. Study the change in carbonic acid in
> the ocean.
Knowing current trend does not mean understanding it and,
more importantly, correctly predicting it.
> Stop with this red herring.
Red herring? Then how would you call the speculations that
oceans are about to release vast amounts of accumulated
CO2 and fry the Earth, without independent study
of mechanics of sea-air interface?
> > > If you are here simply to obfuscate the conversation, I merely ask you
> > > to check in with your conscience before proceeding.
> >
> > It is funny. I initiated this particular conversation, at least the most
> > substantial part of it, and now you are accusing me of obfuscaion of
> > my own conversation? It does not sound as a coherent thought from you.
> > If you want to talk to yourself and yourself only, with self-imposed
> > constraints of misconceptions and rigid belief system, it is advisable
> > to check with your local church.
>
> This is just a bit bizarre, frankly. "Your own conversation"?? And
> it's gratuitously rude. Everytime some one throws this "religious" mud
> it just stinks of cheap shots and a poor effort at creating an alibi
> for their own unwillingness to learn.
What is rude is to continuously allude to my intent
to obfuscate the conversation, to engagement in special
tactics of deception, that my arguments are silly,
that I have blanks in general education, cannot learn,
and advice to check with my awareness of a moral
or ethical aspect to my conduct before continuing
my argumentation.
Quite frankly, I am tired of arguing with you guys about
one thing while you are arguing about your own and different
thing. I will reconsider my intent to participate in this group,
but reserve my right to express my unmoderated
opinion about you folks in general usenet groups.
- Alexi
Coby Beck
I said: "It is clearly most sensible to divide it into land, air and
ocean" You are again being 100% intentionally obtuse.
> > .. Personally, I think you are just playing word games.
Guess I was right.
> > original post was very clear in asking how a global flux could be more
> > precisely known than any of its components.
>
...
> If you (or Mr.Tobis and Prof. Arritt) misconstrued my posts
> and jumped into meanigless discussion of "global constraints",
> it is your problem.
In this message,
http://groups.google.com/group/globalchange/msg/ec895e4ab141556a Alexi
said:
"I have raised a question regarding accuracies in estimation of global
CO2 fluxes, which went unanswered. Maybe this group can help us?
....
Now, could you tell me how it is possible to esimate global CO2 uptakes
and inhales "
Your words. It never fails to amaze me how people will deny their own
words knowing full well they are archived for all to see.
> > > Second, if someone
> > > wants to understand the _trend_ in global inventory (which I guess
> > > is the primary purpose of the whole climatology), one has
> > > to study local fluxes
> >
> > Why do you persist in this ridiculous argument? If you want to know
> > the trend in global inventory of atmospheric carbon dioxide just go
> > look at the Mauna Loa records. Study the change in carbonic acid in
> > the ocean.
>
> Knowing current trend does not mean understanding it and,
> more importantly, correctly predicting it.
Of course. But the topic at hand is measuring it.
> > Stop with this red herring.
>
> Red herring? Then how would you call the speculations that
> oceans are about to release vast amounts of accumulated
> CO2 and fry the Earth, without independent study
> of mechanics of sea-air interface?
I would call it another red herring.
> Quite frankly, I am tired of arguing with you guys about
Good, because quite frankly I don't think I could write another
response without violating our own stated moderation policy.
Coby
Alexi Tekhasski
> On 6/17/06, Alexi Tekhasski <apredte...@austin.rr.com> wrote:
>
> > if someone
> > wants to understand the _trend_ in global inventory
>
> Are you suggesting that we don't understand the trend to first order?
Yes. You have no reasonable explanation yet about natural
oscillations between glaciations and deglaciations. Without
understanding of these natural oscillations you cannot correctly
estimate the response of globe to perturbations.
> > (which I guess
> > is the primary purpose of the whole climatology),
>
> How silly.
>
> No, it's pretty much an established result in *geochemistry*, and
> isn't really a *climatological* result at all. You may wish to gain
> more understanding of fields which you are so quick to criticize.
>
> > [to understand the _trend_ in global inventory] ...
> > one has
> > to study local fluxes.
>
> I think this is trivially and obviously incorrect, for reasons which
> have been explained in great detail.
And I just have outlined the concept that shows that
all your reasons have no practical sense.
> To clarify what it is we are discussing, do you agree that we can be
> confident that the most recent part of this signal is anthropogenic?
>
> http://www.globalwarmingart.com/wiki/Image:Carbon_Dioxide_400kyr_Rev.png
>
I cannot agree with your assertion with sufficient confidence.
The chart looks like the whole historical data are miscalibrated
by a constant shift. I have never seen any analysis of trapped
air method with regard to errors. The method was not tested
with direct controlled experiments AFAIK. There is an opinion
that different dynamics of decompositions of differnt clathrates
leads to substantial underestimation of CO2 concentrations
in the analyzed air samples.
> Do you contend that no global evidence could possibly suffice for such
> attribution?
Whether the attribution is correct or not, it is not relevant.
What is relavant is that you cannot correctly predict dynamics
of the system just looking at global concentration of CO2.
- Alexi
Alexi Tekhasski
I suggest you re-read again what you wrote yourself. You said:
"it is ridiculous to consider the ocean as a "regional" component".
If you might be aware, oceans occupy a continuous set of regions,
and therefore can be combined into one "region", a concept that
is frequently employed in simplified "box" models, and I don't
find any ridiculous content in this approach.
> > > original post was very clear in asking how a global flux could be more
> > > precisely known than any of its components.
> >
> > If you (or Mr.Tobis and Prof. Arritt) misconstrued my posts
> > and jumped into meanigless discussion of "global constraints",
> > it is your problem.
>
> In this message,
> http://groups.google.com/group/globalchange/msg/ec895e4ab141556a
> Alexi said:
> "I have raised a question regarding accuracies in estimation of global
> CO2 fluxes, which went unanswered. Maybe this group can help us?
> ....
> Now, could you tell me how it is possible to esimate global CO2 uptakes
> and inhales "
>
> Your words. It never fails to amaze me how people will deny their own
> words knowing full well they are archived for all to see.
It never fails to amaze me how a person who lost an argument
would resort to quoting out of context and insist on perfect
use of terminology in initial question. If you examine the whole
referenced above text, you would find explicit references to
(point by point of the original message):
(1) "..ocean surface albedo";
(3) "piston velocity";
(4) "..over whole sea area";
Then the post mentions:
"CO2 uptakes and inhales at about 100Gt/y"...
"to arrive at a small difference (ocean uptake) of 2GT/y"
Quite frankly, only a completely ignorant or dishonest person
would confuse the given numbers of 100Gt/y and 2GT/y or
attribute the "piston velocity" to something else other than
oceans, and continue to insist that the topic was about
"totally global fluxes" even if I gave explicit clarifications
in many subsequent posts that it was solely about the
ocean.
The whole example was meant to illustrate the
root of your confusion in estimating CO2
flux over lands, since there are no methods to
measure that flux directly.
> > Quite frankly, I am tired of arguing with you guys about
>
> Good, because quite frankly I don't think I could write another
> response without violating our own stated moderation policy.
Please go ahead, I hope that all thread is archived in Google
for entertainment of future generations. Therefore, I strongly
suggest for you to think of a good exit strategy from this debate
and save some integrity. Keep in mind that I maintain my
own achive, just in case of out-of context quotations and
severe "moderation".
Take care,
- Alexi
Jim Torson
>It never fails to amaze me how a person who lost an argument
>would resort to quoting out of context and insist on perfect
>use of terminology in initial question.
>Please go ahead, I hope that all thread is archived in Google
>for entertainment of future generations. Therefore, I strongly
>suggest for you to think of a good exit strategy from this debate
>and save some integrity. Keep in mind that I maintain my
>own achive, just in case of out-of context quotations and
>severe "moderation".
I would suggest that this list is not a place to "win" or "lose"
arguments. It should be a place for discussions to improve
understanding of issues.
I doubt that anyone cares whether or not you maintain
your own archive to try to "prove" that you "won" an
argument.
Jim
Michael Tobis
> Keep in mind that I maintain my
> own achive, just in case of out-of context quotations and
> severe "moderation".
In turn, you may keep in mind that our moderation policy, while we
actually hope it is constructive, makes no effort to defend its
fairness. Our policy is:
"Posts will be admitted to the list if and only if any moderator finds
the submission to be constructive and/or interesting, on topic, and
not gratuitously rude."
I think one of our moderators is being extremely generous in the
present case. It isn't myself.
People who doubt that there is a legitimate scientific basis for
global change issues are cordially invited to choose one of the
numerous alternative venues to make such a case. I for one do not
consider such a position constructive or even polite.
mt