Estimating tephra fall volume from point-referenced thickness measurements
J Rougier, R. S. J. Sparks, Willy Aspinall, Susan H. Mahony
Abstract
SUMMARY We describe an algorithm for producing an approximate 95 per cent confidence interval (CI) for the bulk volume of tephra fall deposits, based on point-referenced tephra thickness measurements, and expert judgement. Applied to the Aso-4 eruption, our algorithm gives a 95 per cent CI of $[220\, \mathrm{km^3}, 370\, \mathrm{km^3}]$. One component of our algorithm is an estimated thickness surface, from which is it also possible to produce isopach maps, thickness transects and ‘Pyleograms’. Our approach uses modern methods from statistics and machine learning to model and fit the thickness surface, and to produce the 95 per cent CI, and we contrast these with established approaches.