Statistics can be used in volcanology to assist in the analysis of the distribution of volcanic centres in space or events in time. A sequence of explosive events can be represented as a time series at modelled in this fashion. The distribution of the time between events or repose period can be analysed and used to model the activity.
Many studies have been performed with data representing large eruptive events, which tend to be relatively rare. This project is applying statistics to the frequent small explosive eruptions that have been occurring at Volcán de Colima. Many thousands of events have occurred since 2003. Statistics have shown the presence of two event types, which have a distinct seismic waveform. Comparison have also been made with several other volcanoes: Tungaruhua , Ecuador ; Karymsky , Russia ; Erebus, Antarctica and Gunung Semeru , Indonesia .
Treating the data as a time series, statistical models have been fitted using autocorrelation methods. The fractal dimension has been calculated to quantify the degree of clustering within the data. It has been found that the distribution that best fits the repose interval data varies for different eruptive periods. This implies different types of renewal process, which has implications for the dynamics of the eruption.
Dr. Jeff Johnson, University of New Hampshire, USA
Dr. Chuck Connor, University of South Florida