Experimental Turbidite Lamination in a Circular Flume

Experiments on currents in which processes of long duration are involved, e.g., on slowly depositing, eroding, or decelerating suspensions, cannot be carried out in normal, straight flumes. Einstein and Krone therefore used a channel with a narrower and faster return flow. The disadvantage is the turbulent mixing during this return, but they nevertheless obtained noteworthy results. I employ a circular flume with vertical paddles moving horizontally in a circle along the axis. Centrifugal force and differences in velocity along the inner and outer wall are troublesome, but turbulence due to the drive is uniform and slight. The experiments described here are concerned with the origin of horizontal lamination as shown by nearly all turbidites of finer grain than coarse sand. The origin has been attributed to various causes, but usually to current pulsations of different kinds. Lamination is quoted as evidence for, but also against, deposition by a turbidity current. However, current pulsations are so numerous that they should produce ten to a hundred times more laminae than are present. It is shown experimentally that a perfectly steady, decelerating flow, comparable to the lower part of a turbidity current, produces grading, rippling, and convolution. But it also creates lamination, provided the accumulation is not too swift. Adjacent laminae can have median diameters differing by a factor of 3, and brick particles show concentrations by a factor of 6 in some laminae. The selective concentration is due to the tendency of particles moving along the bottom to join stationary ones of equal weight, density, and shape. The latter repel all grains that differ in any of these properties. The principle of "kind seeks kind" dominates the accumulation of grains that are rolling over the bed between excursions into the saltating carpet. An attempt is made to account for this behavior. Horizontal lamination in deposits formed by rivers is believed to result in most cases from the same selective action and not from pulsations. This picture is very close to that deduced earlier by Moss and argued so convincingly on the basis of composition of natural laminae. The absence of lamination in the coarser parts of turbidites is ascribed by Sanders to a postulated new type of sand flow. Here the absence of coarse lamination is attributed to the lack of time for sorting due to swift accumulation (estimated at about 1 cm/min) and to the nature of the movement in a traction carpet. The behavior of the latter is described as observable in a true experimental turbidity current. This is in general accordance with the views of Hsu and of Dzulynski and Sanders on such a carpet. The carpet is a secondary phenomenon dependent on the turbidity current and forming a minor part of it.