Global frequency of large magnitude explosive volcanic eruptions recorded in the last 100,000 years

Hobbs, L.K.

Abstract

The effects of large magnitude (M ≥4.0) explosive eruptions can be catastrophic on a local and potentially global scale. The largest eruptions (M ≥8.0) can disperse ejecta over thousands of km2. As the last known event of this kind took place ~26,500 years ago, the behaviour, and the extent of the possible consequences, of such eruptions are not fully known.
A database was constructed of large magnitude explosive eruptions recorded in the last 1.8 Ma. This contains information on the source volcano, geographical location, age, tephra and dense rock equivalent (DRE) volumes, column height, magnitude, intensity, Volcanic Explosivity Index (VEI) and magma composition, where known, for each eruption. Such an extensive dataset has not been compiled previously. A major obstacle when collecting such data is under-recording of past volcanic activity. Historical and geological records decline in completeness and quality further back in time and with decreasing eruption size. This underrecording must therefore be addressed when analysing past volcanic behaviour.
These data were analysed to investigate: a) the quality of the compiled records, b) the relationship of magnitude with VEI and intensity, c) average return periods of large explosive eruptions and the maximum eruption magnitude possible on Earth and d) the extent of underrecording through time as a function of magnitude.
The accuracy of date and erupted volume data is determined by the methods used to obtain them. The more information that can be found for an event to corroborate size estimates, the more reliable its record is considered. Confidence indices were developed to indicate reliability of age, volume and overall eruption data for each event in the database. These levels allowed filtering of data for analysis and provide a system to assess the quality of the data.
Assessment of the relationship of magnitude with VEI and intensity showed: widespread approximation of magnitude; that the range of magnitudes recorded at each VEI is approximated by VEI ≈ M(X) ± 1; and a limit of intensity exists for large explosive eruptions.
Return periods, predicted magnitude limits and under-recording were analysed using an extreme value statistics threshold exceedance model. The choice of threshold is important as it influences modelled results. Data from 0-100 ka were used as older records were too sparse to produce reliable results. The predicted magnitude limit of 11.7 is larger than the largest known eruption in Earth history (M 9.2) but the average return period for eruptions of M ≥8.0 – 13,274 years – is too short compared to the reasonably complete geological record for these extremely large events. This suggests that too few data were considered, which is supported by a very low prediction of ~30% probability of an M 8.0 event being recorded prior to 50 ka.
However, analysis of the more complete 50 ka record produced an unreasonable return period of 13.9 Ga for M ≥9.2 events. Therefore future improvement should focus on the >50 ka data.

Citation

Hobbs, L. Global frequency of large magnitude explosive volcanic eruptions recorded in the last 100,000 years. (Dissertation). University of Bristol. Retrieved from https://uwe-repository.worktribe.com/output/1019147

• Thesis Type: Dissertation
• Public URL: https://uwe-repository.worktribe.com/output/1019147
• Award Date: Jan 1, 2008