March 14, 2019 | David F. Coppedge

New Waterfall Theory Affects Interpretations of Climate, Tectonics

How do waterfalls form? Geologists tested them in the lab, and found that some of them can make themselves.

It has been long assumed that you need some external event to form a waterfall: a change in rock, a change in flow, a landslide, or some other “allogenic” source. Three scientists, publishing in Nature this week,1 were intrigued that some waterfalls, often formed in a series of drops, occur without any clear change in lithology (rock type). Sometimes they form in unperturbed bedrock.

They built a flume in the lab simulating conditions for heavy water flow in a uniform substrate, and found that waterfalls can make themselves. These “self-formed bedrock waterfalls” occur when feedback between the surface and the rate of flow causes the formation of cyclic pools that deepen until a sheer drop is formed. When that occurs, most of the erosion is at the bottom of the pool rather than farther downstream. One would think this mechanism has been understood for a long time, but that is incorrect:

For example, experiments have shown that incising, steep riverbeds rapidly develop repeating, undulating chutes and pools, known as cyclic steps, that are self-organized. Cyclic steps are similar to waterfalls, but their step height is small relative to their wavelength, and they lack a freefalling jet that defines a waterfall. It has been proposed that cyclic steps might develop into waterfalls, but this hypothesis has not been tested. Instead, previous experiments have explored waterfall formation due to perturbations in external forcing—such as changes in base level, water discharge or sediment supply—or have used a layered or fractured substrate to promote waterfall formation. Moreover, most previous experiments used sediment substrates erodible by clear water alone, whereas natural bedrock channels wear by abrasion from particle impacts or plucking. Techniques using scaled concrete or foam produce better quantitative analogues to bedrock, but experiments using these bedrock analogues have yet to explore waterfall formation.

This time, they added abrading particles to the water, and had it flow over an artificial “bedrock” of polyurethane foam. When the flow reached supercritical rates, cyclic steps, then waterfalls, formed within 2 to 3 hours. In fact, a slot canyon began forming within the first hour, and deepened over the next hour. Using various assumptions, they figure the experimental time would correspond to 100 to 100,000 years in nature, but nobody was present to observe anomalous conditions that might drastically alter their estimates.

Muddy water scours a series of falls in Whitney Canyon, Calif. (DFC)


The new theory appears to indicate that self-forming waterfalls can form relatively rapidly, even in bedrock, and persist for thousands of years. This has implications for geological theories of climate and tectonics. They point out why previous efforts at interpretation have been difficult:

Isolating the relative controls of climate, tectonics, lithology and autogenic dynamics on waterfall formation is challenging because bedrock riverbeds evolve over a thousand years or more and they are subject to autogenic dynamics and allogenic perturbations over varied timescales. To overcome this challenge, we used a physical experiment to test the hypothesis that bedrock waterfalls can form due to autogenic dynamics alone, in the absence of perturbations in external forcing or lithologic controls.

What effect could their lab experiments have on other geological theories?

The existence of knickzones and waterfalls is often used to define landscape topographic disequilibrium with respect to external forcing. By contrast, our experiment suggests that waterfalls may be able to emerge in landscapes with uniform lithology undergoing steady external forcing, which could lead to erroneous identification of changes in climate and tectonics from channel profile inversion. Autogenic waterfalls that emerge in response to externally forced river steepening may change the rates that allogenic perturbations propagate through river networks, further obscuring climate and tectonic signals.

Bridalveil Fall, Yosemite (DFC)

Not all waterfalls are autogenic (self-formed). The retreat of glaciers, for instance, allowed for Yosemite’s famous Bridalveil Fall to form by allogenic (other-cause) processes. Up till now, geologists had assumed that waterfalls could be “used to reconstruct perturbations in environmental forcing, such as climate and tectonism, over millions of years.” When interpreting a landscape from now on, geologists can’t assume that waterfalls tell a story of evolution how climate and tectonics acted over long ages. In uniform conditions at high flow rates, some waterfalls can form relatively quickly without any environmental forcing.

1. Self-formed bedrock waterfalls, Joel S. Scheingross, Michael P. Lamb & Brian M. Fuller. Nature, volume 567, pages 229–233 (13 Mar 2019).

We see again that geological interpretation can lead to moyboy conclusions that are based on untested assumptions. Another question that should arise from the fact that waterfalls can self-generate in bedrock in thousands of years is, do we really need millions of years to explain waterfalls? Retreating waters from a global flood could have created many waterfalls in a matter of days.

This study also teaches another lesson: what other geological theories could be overturned by actual lab experiments? Even this study is not the final word. How similar is a concentrated flow over polyurethane to a large waterfall on granite? No one could run a lab experiment on the scale of Yosemite Falls. And even if they could, conditions today could be different from when a large waterfall actually formed, because nobody was there.

Before swallowing the confident dates listed on national park signs, remember this story and others we have reported that call into question the assumptions on which those long ages are based.


(Visited 545 times, 1 visits today)

Leave a Reply