Space-time correlations in the 1D Directed Stochastic Sandpile model

Sandpile models are known to resist exact results. In this direction, space-time correlations between avalanches have proven to be especially difficult to access. One of the main obstacle to do so comes from taking memory effects in a systematic way along the computation. In this paper, we partially fill this gap and derive recursive relations for the particle filling and avalanche 2-points correlation function in the 1D Directed Stochastic Sandpile. These expressions allow to characterize the sign of the correlations and estimates are provided in the particle filling case. In fact, density correlations are shown to be positively correlated. This behavior is directly related to persistence of the local particle filling. On the other hand, we show that avalanches are anticorrelated in the model. This is interpreted by the fact that avalanches disrupt the system and the damage can only be fully compensated after injecting a sufficiently high number of particles. These results indicate an underlying trade off, between static and dynamic observable, for the system to sit in its stationary state. It appears that this balance is controlled by the conservation of the particle number along the avalanches.

💡 Research Summary

This paper presents a rigorous analytical study of space-time correlations in the 1D Directed Stochastic Sandpile (DSS) model, a prototypical system exhibiting self-organized criticality (SOC). The primary challenge addressed is the characterization of memory effects and correlations between successive avalanches, which have been largely inaccessible in exact treatments of sandpile models.

The authors begin by formally defining the DSS model: a one-dimensional chain of sites where each site holds a number of stable (z) and waiting (w) particles. The stochastic evolution rules dictate that a waiting particle on a site with a stable particle (z=1) must create another waiting particle on the left, while on an empty site (z=0), it can either stabilize (with probability p) or jump to the right (with probability q). In the driven-dissipative setting (particle injection at site 1, dissipation at site L+1), the stationary state is exactly known to be a simple product state where each site is independently in the state (p|1⟩+q|0⟩).

The cornerstone of the analysis is a mapping between avalanche dynamics and a one-dimensional random walk (RW) problem with an absorbing wall. An avalanche starting from the stationary state is equivalent to a symmetric random walker starting at an initial height τ (related to the initial stable configuration) and taking steps in a “time” coordinate corresponding to the spatial site index. The walker moves with probabilities α+ = α- = pq for up/down steps and β = 1-2pq for no step, and the avalanche terminates when the walker’s height reaches zero. This mapping allows the authors to express various avalanche statistics in terms of RW path probabilities. Using the method of images, they provide solutions for the propagator (P(y|\tau, x)) and confirm known results, such as the avalanche size distribution (P(s) ~ s^{-4/3} f(s/L^{3/2})).

Leveraging this RW framework, the paper’s main achievement is the derivation of exact recursive relations for two key two-point correlation functions:

1. Particle filling (density) correlation: The correlation between the stable particle occupation (z) at two different points in space and/or time. The analysis demonstrates that these density correlations are always positive. This is interpreted as a manifestation of “persistence,” where a local particle density tends to remain unchanged over a period.
2. Avalanche size correlation: The correlation between the sizes (s) of two successive avalanches. The derived recursions prove that these dynamic correlations are negative (anticorrelated). The interpretation is that a large avalanche disrupts the system, depleting available resources and making a subsequent large avalanche less likely until sufficient particles are re-injected to “repair” the damage.

The discovery of opposite signs for static (density) and dynamic (avalanche) correlations reveals a fundamental trade-off in maintaining the system’s stationary SOC state. The system balances the persistence of its local static configuration (positive density correlation promoting stability) against the disruptive, resetting nature of its dynamic avalanches (negative avalanche correlation promoting instability). The authors argue that this balance is inherently controlled by the conservation of particle number during avalanche propagation. Additionally, the paper notes that for scalar observables depending only on the stable configuration, time correlations vanish exactly after a finite time τ ≥ L, a property shared with related models like the Oslo model.

In summary, this work provides a rare exact handle on space-time correlations in an SOC model. By cleverly mapping the problem to a random walk and deriving recursive structures, it uncovers a subtle interplay between positive persistence in local states and negative feedback between dynamical events, offering deeper insight into the microscopic mechanisms sustaining self-organized criticality.