34 
5.3 Static stability of the gridded data 
Hydrostatic instability of the gridded profiles poses a problem for a number of oceanographic 
applications. Thus, Jacket and McDougall (1995) had to artificially ’’stabilise” the WOA94 gridded 
data set in order to make it suitable for labelling observed hydrographic profiles with the neutral 
density y" variable. A test for hydrostatic stability was performed on the composite data-set 
used in this study. For the validated historical data 5% of the bottle pairs were found to be 
unstable, whereas for the original (not-validated) WOCE data the percentage was only 1%, 
indicating that inversions are mostly due to measurement errors and should not be 
considered real. Jacket and McDougall (1995) report instabilities of the gridded WOA94 
climatology (see their Fig. 1a, page 382) and note that the problem areas are largely the 
Southern and Arctic Oceans, although all oceans are affected to some degree. We checked 
the latest WOA01 gridded climatology for hydrostatic instability and found, that the problem 
still remains (Fig. 23). 
Instabilities occur at about half of all grid points, with gridded profiles exhibiting more 
unstable standard level pairs in the Polar Ocean, Southern Ocean, Mediterranean Sea, 
Okhotsk Sea and some other regions. The vertical distribution of unstable level pairs and of 
the number of inversions is given in Fig. 24. 
Steric Height Anomaly 500-2500 m 
-180 -160 -140 -120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 
Longitude