Generation of Large Geochemical Data Sets for Single Units of Volcanic Rock: Application of Portable XRF Spectrometry to Zoned Ignimbrites

点击翻译按钮获取中文摘要

High-quality chemical analyses of Earth materials are dominantly conducted in fixed-site laboratories and involve lengthy sample preparation procedures. However, recent technical advances in instrumentation hold the promise of a fundamental shift in geochemical practice. The advent of reliable, high-precision, field-portable analytical instruments is opening a new era in which an increasing proportion of data from volcanic and other rocks will be obtained in the field in real time or near-real time (for example, a fully functional lab could be set up at a field base camp), and high-quality rock analyses can be obtained ever more cheaply and quickly. This project will deploy a portable spectrometer to address an outstanding problem in volcanology that has defied conventional geochemical analysis due to the large number of data points needed, namely the behavior of magma during super-eruptions. Since the 1980s, numerous models and simulations of large volcanic eruptions have been constructed, but their predictive value is largely untested because there is a critical shortfall of data from the products of past super-eruptions. Analysis of a sufficient number of samples in conventional labs simply takes too long and is too expensive. This project will be a test case for the technology, methodology and utility of increasing the number of chemical analyses from the products of a single super-eruption by an order of magnitude or more (thousands rather than tens to hundreds of samples) over current practice. Understanding the compositional structure of zoned silicic magma systems is hampered by a lack of geochemical data. The problem is not that of establishing the range of compositions that are present among eruptive products, nor of investigating the ultimate origins of the magmas; existing data sets are probably already sufficient to address those questions. Rather, the issue is one of analyzing enough samples at numerous locations in order to obtain a statistically valid picture of the tuff?s compositional architecture. The number of analyses required is too large (1,000) to feasibly accomplish using conventional methods such as wavelength-dispersive X-ray fluoresence, and in any case, a complete analysis of each sample is not needed as long as the range of compositions present has been established by conventional methods on a smaller number of samples (~100). Portable X-ray fluorescence (PXRF) technology has advanced to the point where it is ideally suited to such an investigation, because for some elements, especially the critical trace elements Rb, Sr, Y, Zr, Nb and a few others which typically exhibit large variations in zoned rhyolitic tuffs, the precision and accuracy approaches that of full-size wavelength-dispersive XRF. Using PXRF, thousands of analyses can be obtained at very low cost during the course of a two-year research project, starting during fieldwork. This project will acquire a PXRF instrument and develop methods for zoned ignimbrites by carrying out a case study on the 1.61 Ma Otowi Member of the Bandelier Tuff, Valles caldera, NM. On the basis of existing data, overall compositional variations in this unit are sufficiently understood to enable a few elements to be used as proxies for the whole composition. The research will focus on the incompatible elements Zn, Rb, Y and Nb, which exhibit 3-fold to 4-fold variations from early-erupted to late-erupted tuff, and are present at sufficient concentrations to enable high-precision determinations by PXRF. With a large enough data set, a 'sample' of tuff (e.g. many pumice clasts collected from within a 1 m vertical range at a single location) can be described by the distribution of compositions within it, rather than by a single data point as is currently the case. This will then provide a basis for interpreting compositional patterns in the tuff in terms of eruptive and depositional processes, with the goal of 'putting the magma back into the chamber' to arrive at a model of zoning that is more quantitatively constrained than is possible with a conventional data set.