2009-2011, the International
group on tephrochronology and volcanism (INTAV) of
Union for Quaternary Research (INQUA) conducting a
broad, international intercomparison of laboratories
that regularly use electron-beam techniques (both
SEM-EDS and electron microprobe) to characterize and
identify tephra deposits. The intercomparison addressed
multiple goals, especially objective 2, of the INTREPID
project - i.e. to advance understanding and efficacy in
tephra fingerprinting and correlating and to evaluate
and quantify uncertainty in tephrochronology. In an
overall sense, the intercomparison was intended to see
how well the tephra community was doing, similar to what
the radiocarbon community does on a regular basis. By
supplying four well-characterized reference glasses to
approximately twenty participating labs, this project
also widely distributed a uniform set of reference
samples that will remain useful well into the future.
Procedures for this effort were modeled largely after
those of the ongoing G-Probe
effort of the International
Association of Geoanalysts.
In many cases, geochemical differences between tephras can be subtle, requiring high levels of precision and accuracy for successful and correct identification. Fine grain size, instability of glass during analysis (i.e. Na-loss), presence of microcrysts, and sample heterogeneity provide additional challenges. Also, the use of published data is often complicated by small differences between results produced by different laboratories. Analyses of common reference materials are necessary to quantify these differences and thereby provide for more robust tephra correlations. Proficiency testing using common samples is also one of the most effective ways for a laboratory to monitor its performance against both its own expectations and the standards of performance set by other laboratories participating in the same test. Proficiency testing represents, therefore, an 'external' form of quality control that helps to highlight not only reliable determinations made by a laboratory, but also measurements that may be subject to unsuspected bias.