For many years it has been known that there were furrows cutting across the shelf in certain regions, but only since it became possible to obtain large numbers of accurately located soundings of the shelf and slope were such features found to be numerous and widespread. Variously termed canyons, valleys, mock-valleys, and gullies by different authors, they have stimulated a great deal of interest among geologists, and a large literature has been built up dealing with the character and mode of formation of these canyons. The data concerning the topography of the canyons have largely been obtained by national agencies engaged in the careful mapping of nearshore areas. Such data have been used by Veatch and Smith (1939) and by Shepard and Emery (1941) to prepare general and detailed topographic charts of the canyons off the east and west coasts of the United States. Stetson (1936) has carried out independent observations on the east coast, and Shepard
Although the terms listed above have been used more or less synonymously, the size and general character of the canyons vary greatly. Some of them off the mouths of rivers, such as the Hudson (fig. 9), Congo, and Indus Canyons, have depressions that can be traced across the shelf and even into the mouths of the rivers. Some canyons extend across the shelf, but others—for example, many of those shown in the charts prepared by Veatch and Smith—are limited to gashes in the continental slope and do not cut far across the shelf. The upper parts of the canyons are generally found to be steep-walled, V-shaped in profile, with the bottom sloping continuously outward (fig. 8). Some are winding, and many show a dendritic pattern, having smaller tributary canyons. In size they vary from small gullies to vast structures of the same dimensions as the Grand Canyon of the Colorado River (fig. 8).
Profiles of submarine canyons. (A) Transverse profile of the submarine canyon in Monterey Bay compared to a profile of the Grand Canyon of the Colorado River in Arizona (cf. fig. 10). (B) Transverse profiles of a small, steep-walled canyon off the southern California coast. (C) Longitudinal profiles of the Lydonia Canyon and the adjacent shelf and slope. (D) and (E) Transverse and longitudinal profiles of the Hudson Canyon, showing the relation to the adjacent shelf and slope. The locations of the transverse sections (D) are shown on the longitudinal profile. Note the vertical exaggeration in certain of the diagrams and the differences in horizontal scales (A and B after Shepard, 1938; C, D, and E after Veatch and Smith, 1939).
The steep walls of the canyons are generally free of unconsolidated sediment, and in those canyons where special investigations have been made the walls appear to be generally of sedimentary rock; in a few cases (for example, Monterey Canyon off the California coast, fig. 10) the canyons are cut into granite that is overlain by sedimentary rock. The sediments in the bottom of the canyons are generally coarser than those on the adjacent shelves, and in some of them cobbles and gravel have been found.
The following agencies have been advanced as possible causes for the formation of the canyons:
Erosion by submarine currents. Daly (1936) advanced the theory that “density currents” produced by suspension of much fine-grained sediment may have flowed down the slope and cut the canyons, particularly during intervals of lowered sea level during the glacial periods. Density currents occur in reservoirs, but there is no evidence of their existence in the sea, where the density stratification of the water impedes vertical flow.
Spring sapping. Johnson (1939), in a thorough review of the literature concerning the character and origin of submarine canyons, develops the hypothesis that solution and erosion resulting from the outflow of underground water might contribute to the formation of the canyons.
Mudflows and landslides. Mudflows are known to occur in the canyons (Shepard and Emery, 1941) and are agents which tend to keep the canyons clear of unconsolidated debris, but it is doubtful whether they are capable of eroding rock.
Tsunamis or earthquake waves (p. 544). Bucher (1946) pointed out that most of the currents that might be found in canyons are of relatively low velocity and are therefore incapable of active erosion of rock. As a possible explanation of the submarine origin of the canyons he suggested that the rapid currents associated with earthquake waves set up in the sea by violent seismic motion of the sea bottom might be effective agents.
Subaerial erosion. The five explanations listed above are compatible with the formation of the canyons below the sea surface. Because of their many resemblances to river-cut canyons on land, many investigators, notably Shepard, believe that the canyons must have had a subaerial origin. However, there is no accepted geological theory that would account for the world-wide exposure of the shelf and slope within relatively recent geological time. To overcome this difficulty, Shepard has suggested that during the ice ages the amount of water removed from the ocean and deposited as ice caps may have been much greater than ordinarily believed (p. 25).
Topography of the shelf and slope off part of the eastern coast of the United States showing different types of submarine canyons. The Hudson Canyon can be traced far across the shelf; others, such as the Lydonia, Oceanographer, and Hydrographer Canyons, cut into the outer margin of the shelf, while others are restricted to the slope itself. Depth contours in fathoms. (Simplified from chart in Veatch and Smith, 1939.)
Monterey Canyon off the coast of California.
Shepard (Shepard and Emery, 1941) has carefully evaluated the arguments in favor of and opposed to these various hypotheses concerning the origin of submarine canyons, and he concludes that no single hypothesis yet advanced can account for their characteristic features. Problems also exist concerning the processes which remove the sedimentary debris that must be swept into the canyons from the shelf. Mudflows and transportation by currents are known to be operative, but their effectiveness has not yet been determined.