Experiments

Do any real systems behave like the simple computer model above? Although many natural complex phenomena seem to show power laws, the computer models are such extreme caricatures of the physical situations that though they show SOC, it is unclear whether SOC would occur in real systems similarly tested. That is, one would like to perform controlled experiments on real dynamical systems to test the idea that they are attracted to a critical state. Some initial experiments with real sandpiles showed power-law behaviour but others did not. It was soon realized that the reason was probably because real sand is not like the computer particles: Real sand particles are difficult to stop once they start rolling, so that large avalanches are the norm and the dynamics is dominated by individual motion of grains over long distances. By contrast the computer model in the last section implicitly assumed ideal particles of low inertia (so that little kinetic energy was accumulated as it rolled) with power laws resulting from collective dynamics. Thus one is led to design an experiment with particles that do not readily roll or slip over one another. Long-grained rice was used in an experiment performed at the Univeristy of Oslo as it has low inertia and its large-aspect ratio ensured that it did not roll easily. The results of the experiment did indeed show power laws. The moral of this is that while SOC is indeed a physical phenomena, and not just an artifact of simple computer models, it might not necessarily be an automatic state of all complex systems.