Thread: How does radiocarbon dating work?

Here is something that has always puzzled me about radiocarbon dating. (I have only a layman's understanding of the subject).

The decay rate of the isotope C14 is known. If I can measure today's amount of C14 in some piece of wood, and I know the amount of C14 at the time the plant died, I can determine how long the plant has been dead. So far so good.

My problem is: How can I tell with certainty how much C14 was contained in a tree some 50000 years ago? The amount of C14 in the wood is linked to the percentage of C14 in the carbon of the atmosphere at the time. Apparently, this amount is determined mostly by cosmic radiation, which replenishes the ever decaying supply of C14 on Earth. Radiation has been observed to vary (slightly) over time, even within the short period of a half-century in which we paid attention to it. Isn't it a rather brave extrapolation to claim that the amount of C14 on Earth hasn't significantly changed in the last 50000 years? In addition to that, wouldn't you expect concentrations to vary locally on Earth at any given time, depending on radiation patterns and convective distribution of C14 by weather?

There must be a simple answer to this question, but I have never heard anyone even bring up the issue: What is the basis for the assumption that the initial concentration of C14 in a dying organism was the same 50000 years ago as it is today? How would we estimate errors in radiocarbon dating, considering an uncertainty in past C14 concentrations?

As you probably know, radioactive decay is extremely accurate, but the result from this technique of dating is within a millenium/centuries more than within an exact day! Its a good point but Im guessing that your question is the eperimental limit of this technique.

Yes, the decay rate is very accurate. I am not wondering about the rate of decay but about the initial conditions, i.e. the accuracy of the initial concentration of C14 and the assumptions associated with it. If this is the limit of the method, what do we know about it that allows us to make a reasonable estimate on accuracy?

Are the assumptions made just because we don't have any data (extrapolation) or is there a strong scientific basis for it (the major assumption being: C14 concentration has not significantly changed in the last 50,000 years)? Is the accuracy that we assign to the method (say one millenium), just from a wildly guessed fudge factor, recognizing that we have no way of evaluating our assumptions, or is there more to it?

Google 'calibration of radiocarbon dating'. This may help you to believe the accuracy of the method.

Partially helps. Calibration is good and necessary, but you can't calibrate without data. In essence, calibration is applied to take account of a variation of C14 in the atmosphere as far back as we know it (about a half century). That will make carbon dating more accurate for relatively short predictions. Beyond that we simply don't have the data... that's where I see a problem.

A general question, that applies to all fields of predictive science is: If you are facing a complete lack of data, how do you come up with an error estimate for your method? Assumptions can always be made, but how do you quantify their impact on your solution if you have absolutely no data for the time period you are trying to cover?

I think even carbon dating must rely on alternative methods for calibration. What are these methods?