One obvious way you might think of to estimate times is to look at a feature of erosion or deposition, measure how much erosion or deposition has taken place, measure the rate of erosion or deposition, divide the former measurement by the latter, and derive a length of time as our answer.
In this article we shall look at some of the problems involved in using such procedures as dating methods.
Rates of erosion and timeEdit
As an example, we might look at a wave-cut platform, see how far the waves cut away the cliff per year or per decade, see how extensive the platform is, and so find out how long the waves have been cutting the platform.
Or again, we might look at the depth of a canyon, measure the rate at which it is being cut, and then see how long it would have taken to cut it at that rate.
Of course this sort of reasoning could only work if as in these examples the erosional process only cuts away part of the landscape. We need to measure the depth of the canyon by comparison to the remaining rock; or the erosion caused by the waves by measuring the length of the remaining platform. If we were just looking at a horizontal erosional surface, then even if we knew the rate at which it was being eroded, we could not just by looking at the erosional surface figure out how much material was there originally and has been removed.
Rates of deposition and timeEdit
We could for example look at the sediment deposited at a mouth of a river, figure out how much sediment it transports each year, and see how long the river mouth has been in that location.
Or again, we might look at a volcano, see from historical records how often it erupts, see how large the typical lava flow is, and figure out how long it took to build the volcanic cone.
There are, however, problems with these approaches.
One problem is this: geological events vary in intensity, and the larger they are, the rarer they are. So, for example, geologists will talk of a "ten-year storm", one of an intensity that only occurs one year in ten; a "hundred-year storm", of a magnitude that only happens one year in a hundred; and so on. Similar things may be said of volcanic eruptions, of rivers flooding, of earthquakes, and of pretty much anything else.
Now this presents us with a difficulty. If we look at the rates of erosion and deposition as they are happening now, we may be discounting the largest events. In principle it might be possible (for example) that much or even most of the erosion forming a wave-cut platform is performed by thousand-year storms of a magnitude that we have never actually observed on that stretch of coast. Similarly, much or most of a volcanic cone might have been formed by eruptions of a magnitude that that particular volcano only ever undergoes every hundred thousand years.
Then again, when we look at erosion or sedimentary deposition, we must also consider long-term alterations in climate. A river, for example, eroding its banks and depositing sediment at its mouth may be a mere trickle in a cold dry climate when compared to its rate of flow in a hot moist climate; and as we shall see in the articles on paleoclimatology, climates have indeed undergone long-term variations such as these.
In some cases, deposition may have stopped altogether for a time, forming a paraconformity; and if the paraconformity is brief enough not to cause a significant discontinuity in the faunal succession, we might never know about it; and if there are enough such paraconformities, then we could be missing sizable chunks of time when we measure the thickness of sediment and try to estimate its age.
(Note that the problems we have been discussing do not all operate in the same direction: some would cause us to overestimate durations, and some to underestimate them.)
And one more problem: suppose we wanted to use these techniques to find the age of a fossil in (for example) the Tonto Group, which lies near the base of the Grand Canyon. Naively, we might try looking at the layers of sediment lying above it, estimating how long it took the limestone, sandstone, shale, etc to be deposited, add it all up, and arrive at a figure.
The trouble is that the rocks contain a number of unconformities between the bottom and the top, and the top itself, the Colorado Plateau, is also an eroded surface. Each of these surfaces represents vanished sediment which took a certain amount of time to be deposited which we cannot even estimate, because we don't know how much sediment there was; and which took a certain amount of time to be eroded which we also can't estimate for exactly the same reason.
So even if we managed to overcome all the other problems we have mentioned and produce good ball-park estimates for the length of time it took to produce each rock formation, we would be unable to put a date on the lowest rocks and the fossils they contained. The best this would do for us, even if we overcame all the other problems and got our figures exactly right, is supply us with a minimum date which could be too low by any quantity at all.
For these reasons, nineteenth-century geologists barely attempted to put dates on rocks. The best they could do was say that the Earth was old. How old? Very old. And a geological period (the unit of time corresponding to a system) was long. How long? Very long. That millions of years were involved rather than hundreds or thousands was very obvious to them; but so also was the fact that considerations of erosion and deposition would not permit them to perform absolute dating.