HOW SIZE OF TARGET AVALANCHES INFLUENCES PREDICTION EFFICIENCY

Abstract

Bak, Tang, and Wiesenfeld [Phys. Rev. Lett.59, 381 (1987)] introduced their sand-pile (BTW sand-pile) as the cellular automata coming to their critical state without tuning any inner model parameters. The main model features deal with grains falling slowly onto the two-dimensional lattice and a quick deterministic transport of the superfluous grains to the boundary. The simplest modifications of the BTW sand-pile develop a random transport mechanism instead of a deterministic one. The model transportation of the grains generates avalanches. We find that before the big avalanches the height of the pile increases and the singular grains organize themselves in special clusters. These observations lead to the formal algorithm that predicts the big avalanches in advance with a certain efficiency. However the efficiency for the BTW sand-pile is worse than that for its stochastic modifications.