Quote:
Originally Posted by onon
Take any year and the figures show anthropogenic emissions higher than the rate of co2 rise in the atmosphere. That is without the error rates. I pointed out that taking a single year, or the long 250 year period has problems with accuracy (ie due to the error rate) which is why I proposed using a 10 year period.
|
LOL! Of course, onon, do whatever it takes to make the number fit the conclusion you want to arrive at. It makes no difference to me.
Quote:
|
2006 saw over 28 billion tons of co2 emissions, while co2 rose about 18 billion tons.
|
Man, even when slapped in the face with them you still choose to use these absolute statements. Let me rephrase for you:
2006 saw
possible emissions of 28 billion tons of CO2, while CO2
possibly rose about 18 billion tons.
Quote:
|
So in your example you have emissions being overvalued by 20%, and co2 rise undervalued by 40%.
|
No, I don't have them "overvalued" or "undervalued". The possibility of getting the numbers you choose to prove your preconceived conclusion is equal to the possibility of getting the numbers I chose to prove my preconceived conclusion. The difference is that I understand my conclusion was preconceived, you seem to believe your preconceived conclusion is absolute fact.
Quote:
|
Which is why I suggested using a 10 year period in which the error range cannot "meet" co2 emissions with co2 rise.
|
Precisely my point all along. You want to pick and choose the numbers which will prove your conclusion - post hoc, ergo propter hoc.
Quote:
|
I accept due to inaccuracy that for a particular year it could be possible for these not to be true, but I find it very unlikely that both figures would be off by near max error and in the right direction too.
|
Well, good. I guess our debate is over. However, I find it necessary to point out there is no "right direction" in science - there is only observed direction. Those who have a conclusion to prove believe in "right" directions.
Quote:
|
Okay. But have you taken the 20.76ppm rise figure you've got and applied that to 10 * 22 (10 years of 22gTC)?
|
Of course not. Not only would that be poor math, it would be even worse science. You're arguing we can take one measurement, of a highly complex and wildly variable system, and extrapolate that over a ten year period. That's just irresponsible. It reminds me of the Hockey Stick.
Quote:
20.76 * 7.8 = 162 gTC rise in atmospheric co2
10 * 22 = 220 gTC rise in human emissions
So over that 10 year period NC = 162 - 220
NC = -58
So even though nature could be a net emitter in particular years assuming a overvalue for emissions and an undervalue for co2 rise, over 10 years these errors melt away and it becomes much clearer that nature is a net sink. The yearly errors could be seen as noise on an overall trend.
|
Sure, and if that didn't give you the right numbers you could go out to 11 years, then 12, then 20, then 30 - whatever it takes to prove the conclusion, eh? But that wouldn't be science, would it? That wouldn't be honest and impartial analysis, would it? That wouldn't actually reflect reality, would it? But, hell, what does any of that matter?
Actually, since we're allowed to use whatever math we need to prove our conclusions, you should recognize the amount of human contribution wouldn't be 220Gt CO2 over 10 years. If you're only going to apply the error rate of CR to the front and back ends, why not just apply the error rate of HC to only the front and back ends? That way you can get 68.8Gt C ((7.1*8) + (2*6)), or 253Gt CO2 over 10 years. Yea, that's the ticket!
Honest and impartial analysis would require an error range of statistical relevance plotted against the derived estimated averages. Honest and impartial analysis would determine the true yearly anthropogenic emissions, with a statistically relevant error range. I replicated that as closely as possible by using the numbers reported by your sources. But shit, you can't get grants by confusing people, or telling them it's statistically possible things might not be as dire as you want them to believe. You can't make definitive, absolute statements with honest and impartial analysis.
Quote:
|
Well there is natural background noise, for example co2 cycles up and down during a single year like a sine wave due to plants,seasons and temperatures. This wave is noise over the top of the rising trend, so the differences being natural variation rather than measurement error are very likely.
|
Natural background noise is eliminated with the dual measurements. Both measurements include the noise, so it can't be the reason the measurements differ from one another.
Quote:
|
The table isn't presenting what you think it is. Notice the second part is called Partitioning among reservoirs. What are you partitioning? The human emissions. Therefore if you say there are only 6gTC anthropogenic emissions then you get to distribute 6 across those natural storages in the 2nd part of the table (meeting the contraint range of each one).
|
The table is presenting exactly what I think it is, and I never implied nor wrote it said anything different from what you just wrote. If 6GtC are anthropogenic, then that could be entirely absorbed ("partitioned", if you prefer) by non-atmospheric sources. The total of non-atmospheric sources exceeds 6GtC.
Quote:
|
The items and their uncertainty ranges are not independent. "Partitioning among reservoirs" adds up to 7.1 for a reason - it's exactly what anthropogenic emissions are given as.
|
Post hoc, ergo propter hoc, onon. You're making an assumption about one, if not both, of these sides of the equation. You're implying these researchers did not research how much human emissions were absorbed where. You're implying they simply plugged in whatever numbers they felt would prove their point. I guess you're right about one thing - it's not a coincidence.
Quote:
|
Clearly you cannot take all the maximums from the last list and the minimum from the first, and say they spent $15.5 of the $9 given to them. That does not make sense.
|
Yes, it does. Mathematically, statistically, logically, and intuitively. And if your example has the same 90% confidence level these researchers claimed to have, then you would have to wonder what happened. Did someone give you some more money? Are your estimates off? Did you have more money than you thought? Did you spend less than you thought? Did you get the wrong change back? You would have to go back and re-investigate your data.
Or, you could simply plug in whatever numbers prove your conclusion.
Quote:
|
So if you are going to use the lower figure, its 6.0, not 5.0.
|
Nope, again. The lower figure is 5.0 because we're talking about human emissions, not human changes to tropical land use. Changes to tropical land use aren't emissions, they're a perturbation of the natural ability to absorb emissions.
Quote:
|
But that means you have to distribute 6.0 in the "Partitioning among reservoirs" table, you cannot distribute more than the value you choose for anthropogenic emissions.
|
And I didn't. I merely pointed out that atmospheric absorption wasn't needed in order to deal with all human emissions. There's no requirement that any particular partition receive any particular amount, merely that the confidence level is high within those ranges.
Perhaps this will help you:
Total money given to someone: $10 (9-11)
What they did with it (ie partitioning the amount given to them):
Spent it on candy: $3 (2.5-3.5)
Gave it away: $5 (3-7)
Lost it: $2 (1-3)
Now, I'm 90% sure any of these ranges of numbers represent reality. I have less confidence in any numbers outside that range, but they are possible.
So, I have 90% confidence I gave someone $9. I have 90% confidence they spent $3.50 on candy, and 90% confidence they gave $5.50 of it away. Even if I'm only 1% confident they lost some of it, given the numbers I am 90% confident of it would appear they did not lose any of it.
Mathematically, statistically, logically, and intuitively all of the anthropogenic emissions have a 90% confidence level of being partitioned among non-atmospheric sinks. To make this false, you would have to argue (as you are trying to do) that atmospheric partitioning has a 100% confidence level between 3.1 and 3.5. But that's not what your source claims.