Deterministic Versus Probabilistic Volcano Monitoring: Not “or” But “and”

Abstract

Volcanic eruption forecasting and hazard assessment are multi-disciplinary processes with scientific and social implications. Our limited knowledge and the randomness of the processes behind a volcanic eruption yield the need to quantify uncertainties on volcano dynamics. With deterministic and probabilistic methods for volcanic hazard assessment not always being in agreement, we propose a combined approach that bridges the two schools of thoughts in order to improve future volcano monitoring. Expert elicitation has proven to be an effective way to bind deterministic research within a probabilistic framework aiming to reduce the uncertainties related to any hazard forecast; yet, numerous exercises based on expert elicitation have revealed that the attempt to reduce uncertainties led to the creation of new ones, often unquantifiable, created by human nature and reasoning during stressful situations. Such reasoning ignores the complexity of volcanic processes and the fact that every scenario has a probability to occur. The recent probabilistic methods and tools marry probabilistic and deterministic approaches and lead to unprecedented models. Nevertheless, probabilistic hazard assessment is often misunderstood as not all of the researchers involved have backgrounds in such matters. A probabilistic method cannot stand-alone as it depends on data input obtained by deterministic approaches. We propose that, given the symbiotic relationship between the two methods, a probabilistic framework can play a role of moderator between various deterministic disciplines, thus creating a coherent environment for discussion and debate among seismologists, geodesists, geochemists. This can be achieved by training all scientists involved in hazard assessment, probability theory and data interpretation, while at least one group member objectively uses the information provided to produce the probabilities. Hence, numerical outcomes can be interpreted transparently as they represent the quantification of experts' knowledge and related uncertainties. A probabilistic method that incorporates the joint opinions of a group of multi-disciplinary researchers facilitates a more straightforward way of communicating scientific information to decision-makers. Resumen extendidLa previsión de erupciones volcánicas y la evaluación del peligro son procesos multidisciplinarios, con implicaciones tanto científicas como sociales. Nuestro conocimiento limitado de los procesos detrás de una erupción volcánica y su aleatoriedad genera la necesidad de cuantificar las incertidumbres sobre las dinámicas del volcán y de mejorar la política de la toma de decisiones durante una crisis volcánica. Sabiendo que existe un desacuerdo sobre el uso de métodos determinísticos o probabilísticos durante la evaluación de la peligrosidad volcánica, revisamos ambos métodos y proponemos un enfoque que sirve como puente entre las dos escuelas de pensamiento y que pueda mejorar las capacidades de monitoreo volcánico en el futuro, hacia el reconocimiento en tiempo de m...

Partly Cloudy With a Chance of Lava Flows: Forecasting Volcanic Eruptions in the Twenty‐First Century

“…This latter challenge is critical because, in some regards, a probabilistic forecast cannot be “wrong” unless the probability of an event is incorrectly assigned 0 or 100% (even very unlikely events do occur in nature). It is only after performing many probabilistic forecasts and comparing with actual outcomes that we can assess overall performance (Newhall & Pallister, ; Rouwet et al, ). As probabilistic forecasts become more common, it will be vital to perform such analyses and update our approaches as needed.…”

“…Forecast probabilities are ideally as quantitative as possible, although a large degree of qualitative uncertainty is often also necessary (Decker, ). The uncertainties associated with forecasts may be frustrating, but also make them more valuable than predictions that do not quantify uncertainty or the probabilities of alternative outcomes (e.g., Rouwet et al, ). We turn on the news not for weather predictions but for weather forecasts—probabilistic assessments of future conditions.…”

“…In light of these challenges, it seems self‐evident that forecasts should utilize all available information—from knowledge of physics, to monitoring data, to expert opinion, to observations of past events and analog systems—in an integrative, quantitative framework. Furthermore, such forecasts must necessarily be probabilistic, accounting as much as possible for uncertainties in all constituent elements of the forecast (e.g., Rouwet et al, ).…”

Long-Term Probabilistic Volcanic Hazard Assessment Using Open and Non-open Data: Observations and Current Issues

“…One key difference in deterministic, scenario-based hazard maps is that the associated probability of exceedance is not known (e.g., Esposti Ongaro et al, 2008;Charbonnier and Gertisser, 2012;Capra et al, 2015). Nevertheless, it is also important to acknowledge that probabilistic and deterministic approaches should not be seen as an "either-or choice" (e.g., Newhall and Pallister, 2015;Rouwet et al, 2017): both methods are reciprocally informative and beneficial for the goal of improving volcanic hazard assessments (e.g., Marzocchi et al, 2008;Marzocchi and Bebbington, 2012;Newhall and Pallister, 2015).…”

Application of the probabilistic model BET_UNREST during a volcanic unrest simulation exercise in Dominica, Lesser Antilles

“…Hence, in order to improve its results, all effort of the experts involved in the elicitation should focus on providing reliable input information. It is not a tool that “gives the best and final results,” it should not be seen as a tool to “confirm scientists' opinions,” nor should it be seen as a competing opinion [see Rouwet et al ., ]. It is our opinion that the results obtained with BET_UNREST should be included in the reports transmitted to the authorities as they offer a measure of uncertainty, regardless of the numerical value they have.…”