Granular materials lie somewhere between liquids and solids: while a pile of sand at rest looks solid, when you pour extra sand on it from above, you can either obtain a continuous, liquid-like flow of sand on the surface of the pile, or you can get intermittent flow - sometimes called avalanche dynamics. As the flow of sand itself, very often researchers in the field tend to be segregated into two groups: some of them study the statistics of avalanches, while others study the mechanisms of individual avalanches. We have approached granular matter from both sides.

Avalanches in piles of ball bearings

A recent theoretical contribution which is somehow related to our previous study of avalanches in piles of beads has been recently published by our group. It consists in an earthquake model which shows some level of periodicity in the quake -or avalanche- behavior. This result might be of relevance to explain quasi-periodicity in some reports of real earthquakes. To get the paper, click here.

One of the hottest topics when we talk about earthquakes is prediction --a subject full of debate and passion. While BTW-like earthquake models eliminate any possibility of prediction, our own model mentioned in the last paragraph suggests that it is not completely impossible. In fact, by refining our experiments in piles of ball bearings, we have recently given solid ground to believe that avalanche prediction is indeed possible --provided the exponent in the avalanche size distributions is not -1. The trick is that, instead of trying to predict a big avalanche solely based on the time series of events, we use the temporal evolution of the "average structure" of the pile to attempt prediction. It translates into the earthquake world as "prediction might be possible if, instead of using only the time series of events, you use the temporal series of structural or tensional changes deep in the ground". In the words of H. Jensen, a world authority in Complex Systems from the Imperial College, our experiments "open the possibility for prediction" for real physical systems. Our paper can be found here.

Revolving rivers: a new mechanism of pile formation

A few years ago, we serendipitously discovered the phenomenon of "revolving rivers" in a mysterious sand from "Santa Teresa" (Pinar del Rio province, western Cuba ). It constitutes a new way of formation of a sandpile, consisting in "rivers" of flowing sand that revolve around the (static) pile in one direction or the other, as sand is added from top. Depending on the size of the pile, the rivers are "continuous" or "intermittent". We have been able to explain the observations through a phenomenological theory based on an idealized geometry of the pile and mass conservation, but further work is needed to understand why some types of sand show revolving rivers, and others do not. To get our paper in the subject, click here. A video can be downloaded here (3Mb) or (52 Mb).

Soliton-like uphill bumps in granular flows

A second serendipitous discovery also found in sand from "Santa Teresa" is the existence of soliton-like "bumps" that move against the downhill flow of sand on a two-dimensional sand heap. The bumps can be qualitatively understood in terms of "stop-and-go" traffic models, and even their soliton behavior can be explained through a KdV equation that can be derived from a Saint-Venant-like hydrodynamic formulation of the problem. We are currently looking for further experimental evidence. A video can be downloaded here and the last paper can be download here