Intercomparison of dissolved trace elements at the Bermuda Atlantic Time Series station

• A novel statistical approach was developed to compare trace metal profiles.
• Excellent comparability between sampling systems
• Excellent comparability between different analytical techniques

The international GEOTRACES programme is emphasising the use of the GEOTRACES reference sample programme and the importance of cross-over or baseline stations where all aspects of sample collection, filtration, processing and analytical methods can be compared. These intercomparison efforts are of crucial importance to merge or link data from different origins together. The occupation of the Bermuda Atlantic Time Series (BATS) station (31°45.92′N, 64°04.95′W) by the Netherlands GEOTRACES (GA02; 13 June 2010) as well as US GEOTRACES (GA03; 19–21 November 2011) Atlantic section expeditions provided an intercomparison opportunity to confirm the compatibility of the 2 different sampling systems as well as different analytical techniques used. In order to compare the data from different sampling systems or analytical techniques, a new statistical approach was developed to include the analytical uncertainty. Furthermore, modifications to an existing multi-element technique (Biller and Bruland, 2012) were made that increased the number of elements analysed and that allows the time consuming extractions to be done shipboard. Overall, we show excellent agreement between data generated by different sampling systems and analytical techniques. However, while both sampling systems are capable of collecting uncontaminated samples for all the GEOTRACES key elements, some apparent outliers are present. The intercomparison between the different analytical techniques also reveals subtle differences that would have gone unnoticed if only reference samples would have been used, underlining the importance and power of intercomparison stations. The results show that cross-over station data can be used to assess consistency between datasets if these stations have been carefully planned and analysed in combination with (internal) seawater reference samples to assure intra-dataset consistency.