OK, here's the real problem I have with all of this discussion. I understand that correlation is not causation but the cyclic and time dependent nature of any two processes can be viewed as either one causing the other, directly or inversely.
Look at the graph and imagine that the sine wave is one process in time (e.g. Temperature) and the cosine wave is another process (e.g. CO2 concentration).
Then, depending on what time you choose to start looking you can say that cos leads sin by Pi/2 radians (90 degrees) (starting at time t = 0) or that sin leads cos by 3Pi/2 radians (starting at time t = Pi/2).
Now, if you make the mistake of assuming that one curve causes the other it's easy to show either that global warming affects CO2 or that CO2 causes global warming. Worse, if you examine specific time segments you can come up with any conclusion you like; let's assume the sine wave represents process s and the cosine represents process c, then:
From time 0 to Pi/2: c decreases as s increases.
From time Pi/2to Pi: c decreases as s decreases.
From time Pi to 3Pi/2: c increases as s decreases.
From time 3Pi/2to 2Pi: c increases as s increases.
So if we confound the issue by seeing causality we can induce anything we like including CO2 decreases global warming. Does this make sense?