Megafloods, gradualism, and the birth of geology

Altai Republic, southern Siberia, close to Mongolia. All the mountains around are older than 60 million years and we are about 300 m above the Chuja River. And yet we are stepping on recent (Pleistocene) gravels! These gravels are similar to those in every river bed except for their elevation above the river and for the fact that they show no apparent stratification: good indications that they were deposited in turbulent waters at that height, only 15 thousand years ago. The problem is: considering the valley width and the slope along the river in this area (~0.6%), such high water should have moved faster than 30 m/s, implying a discharge of about 100 times the present Amazon river. This event is getting to be known as the Altai Flood (+ info in this pdf).
No meteorological event could explain such a huge water discharge, particularly in a small catchment like that of the Chuja River. Instead, Russian geomorphologists came out during the 80's with this explanation: some tens of kilometers upstream (here), a glacier blocked the Chuja River for some thousands of years and formed a large lake behind. When the ice barrier collapsed, the sudden release of the lake's water produced a gigantic outburst flood with no historical precedent. This is today the most accepted interpretation of features such as the gravels laying along the flanks of the Katun valley (e.g., Herget, 2009).
This picture is taken from the top of the gravel deposits, nearly 300 m above today's river level. The gravels (note their size of a few cm) are interpreted to mark the upper reaches of the flooding waters when they encountered the hill obstacle and then ran up converting kinetic energy into potential energy. Location map. Other photos of this fieldtrip here.
But wait, doesn't all this sound a bit pre-geological? Religions often depict the Earth as being shaped by large floods and cataclysms following Creation. Is geology now acknowledging some truth in those myths? 

Expected return period of earthquakes
as a function of their magnitude
(example from Myanmar, source).
The longer your record, the more
destructive earthquakes you
register. Not quite a gradual Earth!
The study of the mechanisms shaping landscape started only in the 17th century, when the principles of geology (superposition, original horizontality, and lateral continuity of strata, and the uniformitarianism and gradualism of geological processes) were developed by pioneers like Steno, Hutton, Lyell, etc. They developed all this at a time when the dominant view was creationism: the belief that the Earth was created magically in a matter of days and that its landscape was further shaped by cataclysms thereafter, including floods of planetary scale. Gradualism (the assumption that the earth behaves rather smoothly in the long term) was adopted by the newborn geological community in opposition to creationism, as a prolific tool for scientific progress. Gradualism was successful in explaining the biological and geological evolution of the Earth over millions of years, even if at odds with  the historical knowledge we have about floods, volcanoes or earthquakes.

It is therefore of little surprise that the proliferation of evidence for megafloods in the early 20th century (Burr et al., 2010) was responded to with skepticism if not direct opposition from most of the scientific community. It questioned a principle that was at the very core of geology.

Shorelines of the extinct Lake Bonneville as drawn by
Gilbert in 1890. 
The first scientific study of an outburst flood belongs to (who else) G.K. Gilbert, in his visionary book "Lake Bonneville" (1890). During the Bonneville Flood, a large portion of this 52,000 km2 lake (similar in size to Lake Michigan) was released through the Red Rock Pass in Idaho, 14,500 years ago, flooding all along the Snake and Columbia rivers. Today it is accepted that the spill of the lake triggered a peak water discharge of 1,000,000 m3/s, and the lake level decreased by more than 100 m in just a week. However, in this case there was no glacier providing an ice barrier as in Altai. The barrier here was not made up of ice, but an accumulation of alluvial sediments at the water divide.
Some questions that remain open are: How did this barrier control the flood? Is it possible to quantitatively link the mechanical properties of those deposits to the magnitude of the flood? Is it possible to predict the peak discharge produced by an overflowing lake?
These questions are relevant to the understanding of floods produced by the overflow of lakes formed by landslides, such as in the Hunza valley. Analogue experiments of dam breaching such as in this BBC video from Norway are important to get expertise with these events, but a mathematical framework of the physics involved is still needed before producing reliable quantitative models.

Hopefully I will learn more about this after a few months in USGS-Oregon, and I will share it here. In the meanwhile, enjoy Burr's et al. reference summarizing not only the best known megafloods on Earth, but also what looks like very similar events in Mars.

**Herget, J. (2009). Reconstruction of Pleistocene ice-dammed lake outburst floods in the Altai Mountains, Siberia GSA Special Papers , 484 DOI: 10.1130/0-8137-2386-8.1
**Devon M. Burr , Paul A. Carling , Victor R. Baker (2010). Megaflooding on Earth and Mars Cambridge University Press DOI: 10.1017/CBO9780511635632

Disclaimer: I don’t really know much about megafloods (yet). If you wish to be certain, don’t take my word for it. Opinions expressed here are often not even those of the author. Past blog quality is no guide to present or future performance. This may contain traces of nuts. Oh, and I copied this disclaimer from here, without permission.