The investigation of a river delta and the heads of several nearby submarine canyons in western Mexico produced evidence for rapid changes in the configuration and depth of the nearshore portions of canyon tributaries. General scarcity of data on the rates of submarine canyon formation and the relationship to river discharge should make these results of special interest.

The Rio Balsas, one of Mexico's largest rivers, empties into the ocean near the heads of a large submarine canyon that terminates in the Middle America Trench. One of the distributaries of the Rio Balsas presently is discharging at the head of Cañón de la Necesidad, which is being eroded actively. Two inactive canyons are related to former discharge channels of the river. Their heads lie at some distance from shore and are being filled with sediment. The Cañón de Petacalco, not now receiving sediment directly from a Rio Balsas distributary, has remained active because the shoreline has not retreated far. Until about 100 years ago its head was being filled with fine-grained and highly organic sediments from a nearby rivermouth, while the coarse portion of the sediment supply joined the canyon via a tributary farther seaward. Since then the river has shifted away from this canyon, and the horizontally stratified sediments in the canyon head have been incised as much as 20–30 m, as evidenced by three ¹⁴C dates of organic material exposed in the steep to overhanging canyon walls.

The changes in the shallow portion of the Rio Balsas submarine canyons seem to be related to changes in river discharge pattern, either directly or indirectly. A shifting point source of sediment supply either activates a pre-existing, partly filled canyon, or erodes a new one near the new river mouth, whereas the canyon at the abandoned river mouth is deactivated following retreat of the shoreline.

The heads of the different tributaries form a dendritic pattern in Holocene unconsolidated sediment. Subaerial processes are not involved in the formation of these submarine canyons. Thus, a dendritic pattern of submarine canyons is not necessarily indicative of subaerial erosion.