Noble gases (3He, 4He, 20Ne, 22Ne)The water samples for stable noble gas isotopes (3He, 4He, 20Ne, 22Ne) were stored from the CTD/water bottle system without contact to atmospheric air and preventing air bubbles into gas tight copper tubes, which are clamped of at both sides.In the IUP Bremen noble gas lab the samples were pre-processed with a UHV (ultra high vacuum) gas extraction system. Sample gases are transferred via water vapor into a glass ampoule kept at liquid nitrogen temperature. For analysis of the noble gas isotopes the glass ampoules are connected to a fully automated UHV mass spectrometric system equipped with a two stage cryogenic system. Regularly, the system is calibrated with atmospheric air standards (reproducibility <0.2%). Also measurement of line blanks and linearity are done. The performance of the Bremen facility is described in Sültenfuß et al. (2009).The total errors for the noble gas measurements are estimated to be < 0.5 % for 4He, < 0.5 % for total Ne, and < 0.5% for δ3HeTransient tracer (CFC-12, SF6)Chlorofluorocarbon (CFC-12) and sulfur hexafluoride (SF6) water samples from the CTD-bottle-system were stored in glass ampoules without contact to the atmosphere during the tapping. Immediately after sampling the ampoules are flame sealed after a CFC free headspace of pure nitrogen had been applied.The determination of concentrations in the IUP Bremen gas chromatography lab is accomplished by purge and trap sample pre-treatment followed by gas chromatographic (GC) separation on a capillary column and electron capture detection (ECD). The system is calibrated by analyzing several different volumes of a known standard gas. The loss of trace gas into the headspace is considered by a careful equilibration between liquid and gas phase under controlled conditions before the sealed ampoules are opened and a precise measurement of the volume of the headspace. Concentrations are calibrated on SIO98 scale (Prinn et al., 2000). A more detailed description of the measurement system is given by Bulsiewicz et at. (1998).

The tracer data set was carefully checked for accurate measurements and outliers. According to the WOCE standards the following flags were applied to each measurement:flag 2 = goodflag 3 = doubtfulflag 4 = badflag 6 = mean of replicatesflag 9 = no measurement (then, the data value is missing)