Long- and short-term volcanic hazard assessment of El Chichón Volcano (Mexico) through Bayesian inference

Abstract

The 1982 eruption of El Chichón volcano constitutes the worst volcanic disaster in Mexico producing more than 2000 fatalities, thousands of displaced people and severe economic losses. This eruption took by surprise authorities, population and scientists, thus preventing the implementation of timely and effective mitigation measures. Here, we use a Bayesian inference approach to provide simple, objective and quantitative schemes for long- and short-term hazard assessment for El Chichón volcano. For long-term assessment, we present the event tree for this volcano including the probabilities of different scenarios based on its past activity. For instance, the probability of at least one magmatic/hydromagmatic unrest episode (4.0%) producing an explosive eruption (75%) with VEI 4 (21.4%) in any 10-yt time interval is 0.64%. Moreover, we included additional nodes for the threatened zones and population according to the published hazards maps. For short-term assessment, we use a Bayesian method to examine the evolution of indicators derived from volcano monitoring. As a case study, we apply this method to the available data of the 1982 eruption of El Chichón volcano. Our results show that this method is useful to identify indicators associated with different eruptive phases, recognize significant changes and underscore the lessons from this eruption. Furthermore, this method graphically depicts the evolution of the indicators, easing the communication with non-specialist during volcanic crises. When this highly explosive volcano reactivates again, the methods presented here can be used as a framework to analyze monitoring data and facilitate the implementation of timely mitigation actions.