Abrasion mill analysis of fluvial clast (>4.0 mm) abrasion in a meandering reach of the Rio Grande, south-central Colorado, indicates that (1) abrasion rates of natural clasts from various locations in the channel and flood plain differ very little, (2) abrasion rates in the first few hours are high, then drop to much lower levels for these natural clasts, and (3) a previously experimentally abraded clast sample, otherwise identical to natural samples, had a much lower initial abrasion rate than any natural samples. The initial high abrasion rate of all natural clasts can be attributed to the presence of an easily abraded weathered layer. We interpret the relatively abrupt drop in abrasion rate after a few hours of abrasion to the complete removal of the weathered layer and the exposure of more resistant rock. The low initial abrasion rate of the previously experimentally abraded sample supports this interpretation and also indicates that none of the clasts in the river have been abraded as extensively as the experimentally abraded samples. From this we infer that clasts have not traveled very far in the channel.
These results have led to the development of a weathering-controlled abrasion model to explain the effect of abrasion on downstream fining (size reduction over distance downstream). Weathering-controlled abrasion is essentially a step function in time, composed of the alternation of long periods of weathering with no abrasion (flood-plain storage) with very brief but intense periods of mechanical abrasion (channel transport). This alternation of weathering and abrasion can produce a net abrasion rate sufficient to generate the amount of downstream fining present along the Rio Grande and also can explain why many abrasion mill experiments (which simulate abrasion but not weathering) fail to reproduce natural abrasion rates.
Changes in the amount of fluvial downstream fining in an aggrading basin margin (and thus, in the rock record) can result from variation in selective deposition (temporary or permanent flood-plain storage) or from changes in the abrasion rate. Experiments demonstrate, however, that the complex interaction between these two processes may in some cases obscure the tectonic signal that is commonly interpreted from basin margin depositional geometries.