Extreme events and predictability of catastrophic failure in composite materials and in the Earth

摘要: Despite all attempts to isolate and predict extreme earthquakes, these nearly always occur without obvious warning in real time: fully deterministic earthquake prediction is very much a ‘black swan’. On the other hand engineering-scale samples of rocks composite materials often show clear precursors dynamic failure under controlled conditions laboratory, successful evacuations have occurred before several volcanic eruptions. This may be because earthquakes are not statistically special, being an emergent property process rupture. Nevertheless, probabilistic forecasting event rate above given size, based on tendency cluster space time, can significant skill compared say random failure, even real-time mode. We address questions this debate, using examples from Earth (earthquakes, volcanoes) including following. How we identify ‘characteristic’ events, i.e. beyond power law, model selection (do dragon-kings exist)? do discriminate quantitatively between stationary non-stationary hazard models (is dragon likely come soon)? Does system size (the dragon’s domain) matter? Are there localising signals imminent catastrophic able access effectively invisible approach)? focus effect sampling effects statistical uncertainty identification events their predictability, highlight strong influence scaling time as outstanding issue addressed by quantitative studies, experimentation models.