Power laws statistics of cliff failures, scaling and percolation

TL;DR

This study reveals that cliff failure sizes follow power law distributions with universal exponents, and proposes a scaling relation between these exponents, supported by a minimal numerical model rooted in percolation theory.

Contribution

It introduces a scaling hypothesis linking failure size distributions and geometry, and develops a minimal percolation-based model that reproduces observed statistics.

Findings

Failure sizes follow power law distributions with specific exponents.

A scaling relation between the exponents is proposed and supported.

Numerical simulations align with field measurements, indicating universality.

Abstract

The size of large cliff failures may be described in several ways, for instance considering the horizontal eroded area at the cliff top and the maximum local retreat of the coastline. Field studies suggest that, for large failures, the frequencies of these two quantities decrease as power laws of the respective magnitudes, defining two different decay exponents. Moreover, the horizontal area increases as a power law of the maximum local retreat, identifying a third exponent. Such observation suggests that the geometry of cliff failures are statistically similar for different magnitudes. Power laws are familiar in the physics of critical systems. The corresponding exponents satisfy precise relations and are proven to be universal features, common to very different systems. Following the approach typical of statistical physics, we propose a "scaling hypothesis" resulting in a relation between the three above exponents: there is a precise, mathematical relation between the distributions of magnitudes of erosion events and their geometry. Beyond its theoretical value, such relation could be useful for the validation of field catalogs analysis. Pushing the statistical physics approach further, we develop a numerical model of marine erosion that reproduces the observed failure statistics. Despite the minimality of the model, the exponents resulting from extensive numerical simulations fairly agree with those measured on the field. These results suggest that the mathematical theory of percolation, which lies behind our simple model, can possibly be used as a guide to decipher the physics of rocky coast erosion and could provide precise predictions to the statistics of cliff collapses.