Antropogenic seismicity and the breakdown of the self-similarity described by nonextensive models

A novel model is presented which considers multimodal and multisources approaches and the recent models proposed in the Tsallis formalism that consider memory in the seismic processes. It is shown that the multimodal Sotolongo Costa distribution fits better than current models for antropogenic seismic real data. The value of the entropic parameter q in many cases varies significantly for the different modes in the multimodal Sotolongo Costa distribution. Thus, changes are shown for the correlation structure in the low energy tremors and the high energy ones. The seismic hazard are compared using unimodal and multimodal models. We show that the first models in many cases underestimate the occurrence probability of damage quakes compared to the multimodal models, which can lead to misinterpretations on the evolution of mining-related seismicity.