This raises another question: if modern rates of sediment accumulation on deltas are one meter per year, in 1000 years that should give one kilometer of sediment. What would happen in a million years? Yes, I know, the sediment is reworked into deeper water, etc, but we don't have all that many sequences in the fossil record that are deep water (something that I think is itself long overdue for correction, but that is for another day), so we invoke recycling, which is convenient, except that it eliminates the evidence for evolutionary sequences, which is also convenient, except that it just ain't so. Stratigraphic sequences I have studied in detail(mostly early Paleozoic) show very little evidence for reworking between beds, and certainly lack any substantial evidence for displacement of massive amounts of sediment to deeper water. Hypothesizing that sediment rates were far lower in the early Paleozoic doesn't cut it since generally they are believed to have been much greater, since there was presumably no vegetation to prevent erosion. In fact the whole problem of sedimentation rates through time has been the subject of a lot of study. The whole affair calls into question our current understanding of sedimentary processes. (Schindel, op.cit; Anders, M.h., S.W. Krueger and P.M. Sadler, J. Geol. 1987. 95:1) The conclusions (borrowed from a friend) are worth sharing:
"Sedimentation rates [estimated by every conceivable technique] are alarmingly variable, spanning at least a dozen orders of magnitude. In every environment the rates decrease progressively as longer time spans are considered. The regression is partly a function of environment. There are several consequences of this pattern and they are not all intuitively obvious.
1. Stratigraphic completeness is a function of time scale and can be meaningfully quoted only for a specified time resolution.
2. Short term sedimentation patterns are a disasterous guide to long term accumulation and we should distance ourselves from the doctrine of uniformitarianism.
3. If high resolution is required then stratigraphic sections must generally be expected to be disasterously incomplete.
4. Even abyssal ooze accumulations may be quite incomplete.
5. Beyond time spans of about 10^4 yr all sections at or above marine wave base seem to be controlled by tectonic and not sedimentological factors. Thus, we should not automatically expect a fluvial section to be less complete than a shallow marine section. That we usually do, is an unfortunate consequence of uniformitarianism.
6. Very few, if any sections of sedimentary rocks will yield a record adequate to study species-level biological phenomena, if a fairly complete sequence of generations is needed. The staratigraphic record may not qualify as a testing ground for punctuated, allopatric speciation. It may not ever be able to tell us if an asteroid arrived at the right time to cause a catastrophic extinction at the end of the Mesozoic."