4. Accumulation of metals into the oceans
In 1965, Chemical Oceanography published a list of some metals' "residency times" in the ocean. This calculation was performed by dividing the amount of various metals in the oceans by the rate at which rivers bring the metals into the oceans.
Several creationists have reproduced this table of numbers, claiming that these numbers gave "upper limits" for the age of the oceans (therefore the Earth) because the numbers represented the amount of time that it would take for the oceans to "fill up" to their present level of these various metals from zero.
First, let us examine the results of this "dating method." Most creationist works do not produce all of the numbers, only the ones whose values are "convenient." The following list is more complete:
Al - 100 years Ni - 9,000 years Sb - 350,000 years
Fe - 140 years Co - 18,000 years Mo - 500,000 years
Ti - 160 years Hg - 42,000 years Au - 560,000 years
Cr - 350 years Bi - 45,000 years Ag - 2,100,000 years
Th - 350 years Cu - 50,000 years K - 11,000,000 years
Mn - 1,400 years Ba - 84,000 years Sr - 19,000,000 years
W - 1,000 years Sn - 100,000 years Li - 20,000,000 years
Pb - 2,000 years Zn - 180,000 years Mg - 45,000,000 years
Si - 8,000 years Rb - 270,000 years Na - 260,000,000 years
Now, let us critically examine this method as a method of finding an age for the Earth.
The method ignores known mechanisms which remove metals from the oceans:
Many of the listed metals are in fact known to be at or near equilibrium; that is, the rates for their entering and leaving the ocean are the same to within uncertainty of measurement. (Some of the chemistry of the ocean floor is not well-understood, which unfortunately leaves a fairly large uncertainty.) One cannot derive a date from a process where equilibrium is within the range of uncertainty -- it could go on forever without changing concentration of the ocean.
Even the metals which are not known to be at equilibrium are known to be relatively close to it. I have seen a similar calculation on uranium, failing to note that the uncertainty in the efflux estimate is larger than its distance from equilibrium. To calculate a true upper limit, we must calculate the maximum upper limit, using all values at the appropriate extreme of their measurement uncertainty. We must perform the calculations on the highest possible efflux rate, and the lowest possible influx rate. If equilibrium is within reach of those values, no upper limit on age can be derived.
In addition, even if we knew exactly the rates at which metals were removed from the oceans, and even if these rates did not match the influx rates, these numbers are still wrong. It would probably require solving a differential equation, and any reasonable approximation must "figure in" the efflux rate. Any creationist who presents these values as an "upper limit" has missed this factor entirely. These published values are only "upper limits" when the efflux rate is zero (which is known to be false for all the metals). Any efflux decreases the rate at which the metals build up, invalidating the alleged "limit."
The method simply does not work. Ignoring the three problems above, the results are scattered randomly (five are under 1,000 years; five are 1,000-9,999 years; five are 10,000-99,999 years; six are 100,000-999,999 years; and six are 1,000,000 years or above). Also, the only two results that agree are 350 years, and Aluminum gives 100 years. If this is a valid method, then the age of the Earth must be less than the lowest "upper limit" in the table. Nobody in the debate would agree on a 100-year-old Earth.
These "dating methods" do not actually date anything, which prevents independent confirmation. (Is a 19 million year "limit" [Sr] a "confirmation" of a 42,000 year "limit" [Hg]?) Independent confirmation is very important for dating methods -- scientists generally do not place much confidence in a date that is only computed from a single measurement.
These methods depend on uniformity of a process which is almost certainly not uniform. There is no reason to believe that influx rates have been constant throughout time. There is reason to expect that, due to a relatively large amount of exposed land, today's erosion (and therefore influx) rates are higher than typical past rates.
There is no "check" built into these methods. There is no way to tell if the calculated result is good or not. The best methods used by geologists to perform dating have a built-in check which identifies undatable samples. The only way a creationist can "tell" which of these methods produce bad values is to throw out the results that he doesn't like.
One might wonder why creationist authors have found it worthy of publishing. Yet, it is quite common. This argument also appears in the following creationist literature:
Baker (1976, p. 25)
Brown (1989, p. 16)
Morris (1974, pp. 153-156)
Morris & Parker (1987, pp. 284-284 and 290-291)
Wysong (1976, pp. 162, 163)
Obviously, these are a pretty popular set of "dating" mechanisms; they appear frequently in creationist literature from the 1960s through the late 1980s (and can be found on many creationist web sites even today). They appear in talk.origins more often than any other young-Earth arguments. They are all built upon a distortion of the data.