The 2010 Eyjafjellajökull volcanic ash sampling experiment: factors and conditions affecting field sampling variability
Kim H. Esbensen,a Pentti O. Minkkinenb and Hans S. Møllerc aIndependent researcher, owner KHE Consulting, visiting, affiliated and guest professor at four universities bProfessor emeritus Lappeenranta University of Technology, Finland; owner SIRPEKA consulting cSpecialist; Bioenergy and Thermal Power COWI Consulting, Denmark
In August 2010, a party of Icelandic explorers undertook a sampling experiment of the famous Eyjafjella volcanic eruption local ash falls. The possibility of having 10 two-person sampling groups with very different Theory of Sampling (TOS) competence and experience proved an opportunity too interesting to miss. The Eyjafjellajökull field experiment constitutes a Replication Experiment performed on a lot of significant generic complexity and heterogeneity, illustrating primary sampling when there is absolutely no control over the original lot and its history. In such cases, the effects on the total sampling variance (i.e. the total Measurement Uncertainty) reflects a compound of factors that cannot be fully resolved even after a comprehensive experimental design. The ultimate confounding concerns sampling competence vs lot heterogeneity. The lessons learned have easy-to-interpret educational relevance for many other types of material lots with similar characteristics and heterogeneity in science, technology and industry. It was also a lot of fun.