Determination of the largest clast sizes of tephra deposits for the characterization of explosive eruptions: A study of the IAVCEI commission on tephra hazard modelling

摘要

The distribution of clasts deposited around a volcano during an explosive eruption typically contoured by isopleth maps provides important insights into the associated plume height, wind speed and eruptive style. Nonetheless, a wide range of strategies exists to determine the largest clasts, which can lead to very different results with obvious implications for the characterization of eruptive behaviour of active volcanoes. The IAVCEI Commission on Tephra Hazard Modelling has carried out a dedicated exercise to assess the influence of various strategies on the determination of the largest clasts. Suggestions on the selection of sampling area, collection strategy, choice of clast typologies and clast characterization (i. e. axis measurement and averaging technique) are given, mostly based on a thorough investigation of two outcrops of a Plinian tephra deposit from Cotopaxi volcano (Ecuador) located at different distances from the vent. These include: (1) sampling on a flat paleotopography far from significant slopes to minimize remobilization effects; (2) sampling on specified-horizontal-area sections (with the statistically representative sampling area depending on the outcrop grain size and lithic content); (3) clast characterization based on the geometric mean of its three orthogonal axes with the approximation of the minimum ellipsoid (lithic fragments are better than pumice clasts when present); and (4) use of the method of the 50th percentile of a sample of 20 clasts as the best way to assess the largest clasts. It is also suggested that all data collected for the construction of isopleth maps be made available to the community through the use of a standardized data collection template, to assess the applicability of the new proposed strategy on a large number of deposits and to build a large dataset for the future development and refinement of dispersal models.