Some numerical considerations in the geochemical analysis of distal microtephra

Abstract

The use of Late Quaternary tephra horizons as isochronous markers for high resolution correlation between records is of growing importance in a number of scientific fields. Particular advancements have been made through the identification of microtephra deposits (very distal deposits of volcanic ash invisible to the naked eye). The successful correlation of ash layers and volcanic eruptions depends on the acquisition of the major element chemistries of a tephra layer, derived from microprobe analyses of individual glass shards. This is particularly important for microtephras where aeolian fractionation has removed much of the mineral phase of the ash deposit, which often aids tephra discrimination. There are, however, difficulties in distinguishing precisely between different eruptions of the same volcano using major element composition and this may not always be resolved by the use of trace elements. Given the potential of tephrochronology as a correlative tool in many disciplines it is important to resolve this difficulty. One way forward is to perform more robust statistical analyses on the geochemical data, as in general, analyses are confined to a series of bi-plots of major elements. Here the authors explore some of the problems associated with dealing with microprobe data generated for individual tephras and pay particular attention to the ‘unit sum problem’. Using a subset of data generated as part of a EURODELTA project on tephra distribution in the Adriatic, the authors demonstrate that this problem is detectable in major element data from tephras and have consequently applied the logratio method before further analyses of the data. It is demonstrated that the use of logratios combined with discriminant functions analysis provides a more robust assessment of likely chemical correlations between tephras, and are superior to the use of bi-plots alone; and obviate the need for any data normalization, a particular point of contention between tephrochronologists.