Salinity intercalibration JONSDAP 76

Seven institutes from six countries around the North Sea took 390 salinity samples for a salinity intercalibration exercise during JONSDAP 76. For samples in glass bottles, a mean dispersion in duplicates of 0.03 promille S was found, between salinities of samples determined (soon) after collection at the institutes and salinities of corresponding samples measured after 7-10 months at NIOZ. Evidence points to two major sources of variation: a. inability of several types of glass bottles to preserve salinity during long term storage. b. insufficient rinsing of sample bottles during collection of samples. Systematic differences between institutes and NIOZ may be attributed to: a. long term storage effects which increase salinity and/or b. salinometers giving mainly 0.01-0.03 promille S low values. However one glass bottle serie shows the possibility of low mean dispersion between Institute and NIOZ (0.002 promille S) combined with a low but detectable systematic difference (0.005 promille S). Thin-walled (low density?) polythene bottles are useless for longterm storage of salinity samples. In summary: for nowadays standards and instrumental possibilities disappointing results were obtained. If these results are representative for the precision of salinities obtained during JONSDAP 76, then a large part of these salinities are useless for the construction of reliable salinity distribution charts and for vertical stability calculations. The salinity determination methods must be cured of defects by cautious inspection of all relevant details