Measurements of atmospheric mercury species during an international study of mercury depletion events at Ny-Alesund, Svalbard, spring 2003. How reproducible are our present methods?

Six groups participated in an international study of springtime atmospheric mercury depletion events (AMDEs) at Ny-angstrom lesund in the Norwegian Arctic during April and May 2003 with the aim to compare analytical methods for measurements of atmospheric mercury species and study the physical and chemical processes leading to AMDEs. Five groups participated in the method comparison that was conducted at three different locations within Ny-angstrom lesund. Various automated and manual instrumentation were used to sample, measure and compare gaseous elemental mercury (GEM), reactive gaseous mercury (RGM) and mercury associated with particles (Hg-P). The concentration of GEM was reproducible during background conditions. For the first time using ambient air, the statistics associated with round robin test procedures were applied. This was found to be an appropriate tool to investigate the reproducibility of GEM measurements in ambient air. The precision for each group measuring GEM concentrations was found to be consistently good (within 5%). Five AMDEs were recorded during the study. Using four different methods, including single and replicate samples, all groups recorded higher values of RGM and Hg-P during AMDEs. The results show that measuring comparable atmospheric mercury species at both the same and different locations (within the Ny-angstrom lesund area) is difficult. Not only do site location and site characteristics create challenges when trying to intercompare results but there are difficulties, as well, in obtaining comparable results with similar sampling and analysis methods. Nevertheless, with our current procedures for atmospheric mercury identification we can differentiate with certainty between "high" and "low" concentration values of RGM and Hg-P. (c) 2005 Elsevier Ltd. All rights reserved.