C . REDUCTION OF DATA 



The computer programs of 1971 were modified to record and 

 process data on magnetic tape. This was essential because the expanded 

 depth scale resulted in nearly five times as many data points per lowering. 

 Recording on cards would have required about 50,000 cards. The data 

 are on several files of a single tape with master card information as be- 

 fore, but the water properties are written in format F6.2, 2F6.3, F7.2, 

 F7.4, slightly different from 1971. Tape has its awkard aspects; 

 particularly, minor corrections and rearrangements are much more dif- 

 ficult to do. The final output tape may reflect this; the stations now 

 are not in perfect serial sequence and a few faulty outputs are inter- 

 spersed because it is too time-consuming to remove them. Details 

 of the tape output are given in the appendix. 



The data listings are similar to those in 1971 but only every 

 fifth depth level was listed in order to keep the output to a reasonable 

 bulk. Data is stored on tape at the maximum resolution, 0.0644 m; 

 condensation by any integral depth factor is possible by parameter 

 changes for either listing or card punching. 



The plotted station data usually shows both the shallow and 

 deep lowerings on the same plot. Occasionally, this is not possible 

 where the two lowerings ended up on separate tape files. The shallow 

 and deep lowerings are then presented separately or one is hand-traced 

 upon the other. There was at times an unfortunate tendency to make the 

 shallow lowering as short as possible with the result that there is 

 sometimes a gap between the lowerings. Where the records overlap 

 greatly, the curves may be difficult to separate. Hand-entered lettering 

 is then inserted to assist the reader. Also, the temperature trace is 

 marked with crosses and the salinity with dots every 20 depth incre- 

 ments. The overlaps will eventually make possible a more critical 

 analysis of the behavior of the shunt. 



The effect of the shunt was at first computed according to 

 conductivity tables, based on its effect in air on the apparent salinity 

 of the STD already shunted with another shunt in air. Comparisons 

 were then made on the tabular data from the first fifth of the stations 

 and an empirical correction of +0.3 o/oo was applied to the theoretical 

 equation. This seems reasonable because the wire passing through the 

 core of the cell occupies a fraction of the cross- section and leads to 

 low results in water. A calculation of the resulting area reduction 

 agrees approximately with the empirical correction. However, there 

 has been no opportunity to check this conclusion with the aid of the 

 plots from all the stations . 



47 



