On Selection of Analog Volcanoes
- 163 Downloads
Estimating the occurrence probability of volcanic eruptions with VEI ≥3 is challenging in several aspects, including data scarcity. A suggested approach has been to use a simple model, where eruptions are assumed to follow a Poisson process, augmenting the data used to estimate the eruption onset rate with that from several analog volcanoes. In this model the eruption onset rate is a random variable that follows a gamma distribution, the parameters of which are estimated by an empirical Bayes analysis. The selection of analog volcanoes is an important step that needs to be explicitly considered in this model, as we show that the analysis is not always feasible due to the required over-dispersion in the resulting negative binomial distribution for the numbers of eruptions. We propose a modification to the method which allows for both over-dispersed and under-dispersed data, and permits analog volcanoes to be chosen on other grounds than mathematical tractability.
KeywordsEmpirical Bayes Conway–Maxwell–Poisson process Information gain
Unable to display preview. Download preview PDF.
- Conway RW, Maxwell WI (1961) A queueing model with state dependent service rate. J Ind Eng 12:132–136 Google Scholar
- Siebert L, Simkin T (2002) Volcanoes of the world: an illustrated catalog of holocene volcanoes and their eruptions. Global Volcanism Program Digital Information Series, GVP-3. Smithsonian Institution, http://www.volcano.si.edu/world/. Accessed 15 March 2010
- Simkin T, Seibert L, McClelland L, Bridge D, Newhall C, Latter JH (1981) Volcanoes of the world. Hutchisson Ross, Stroudsburg, 232 Google Scholar
- Shmueli G, Minka TP, Kadane JB, Borle S, Boatwright P (2005) A useful distribution for fitting discrete data: revival of the Conway–Maxwell–Poisson distribution. Appl Stat 54:127–142 Google Scholar
- Wickman FE (1966) Repose-period patterns of volcanoes. Ark Mineral Geol 4:291–367 Google Scholar