All (lain has lieen used in llio analysis presented 

 in this report.) 



Projrrani FIXAV, which input the AVCOR 

 tape, reavera<red the data after data levels wiiicli 

 failed to pass the AVCOR si<rnia-t test were 

 removed. The output of FIXAV is a printout 

 and a computer card deck. The FIXAV print- 

 out was (luality controlled by recheckinjr tiie 

 sif;nia-t values to ascertain the effect of the data 

 level deletions on the FIXAV run. For various 

 reasons, the zero meter depth level is not re- 

 corded by the DDL. Zero level data is obtained 

 from the STD trace or extra|)()lati()n. and en- 

 tered into the computer card deck. 



The fourth piofrram. SIGPT. determiiuMi the 

 standard and si<;nificant levels, whose tempera- 

 ture and salinity would accurately represent the 

 oripinal water cohnnn. Standard levels were 

 taken at the depths fallin<r closest to minimum 

 recorded depth, 10, 20, 30, .50, 75, and 100 meters, 

 every 25 meters to 300 meters, and then every 

 50 meters to 1000 meters. The first test for sijr- 

 nificant levels consisted of fittin<r a cubic curve 

 throujrh five consecutive temperatuie data points. 

 If the curvature at the midpoint exceeded an 

 absolute value of 0.005. the second, third, and 

 fourth points were compared with the data points 

 immediatelv above and below. A level was sisr- 

 nihcant if it departed fiom a straijrht line be- 

 tween the adjacent points by more than 0.04°C 

 for temperature (more than 0.06%„ for salinity). 

 The second test compared the differences between 

 the curvature of two successive midpoints. If 

 the absolute value of the difference exceeded 

 0.005, the departure of the point from the ad- 

 jacent points was af!;ain checked, usinjj the same 

 limits as in the first test to determine if the point 

 was significant. If both of these tests were nega- 

 tive, the departure of levels from points inune- 

 diately above and below was again checked. If 

 the absolute departure was greater than O.On for 

 both temperature and salinity, the level was sig- 

 nificant. If tiie limits were not exceeded in any 

 of the three tests, the level was not significant. 

 After running the same checks for salinity, the 

 top level of the five level group was dropperl and 

 the next new level was added onto the bottom 

 end, and tlie testing was begun again. The out- 

 put of SIGPT was a printout and computer card 

 deck. The printout was checked for obvious 

 errors such as wrong input. 



The final i)rogram, SARCS, plots temperature, 

 salinity, and sigma-t versus depth, and also plots 

 a T-S diagram. The output, in addition to the 

 plots, consi.sts of a printout and computer card 

 deck. The printout was subjected to a final 

 quality control ba.sed on a careful study of the 

 plots which indicated that the data reported 

 heiein was not grossly unreasonable. The card 

 deck was submitted to XODC. (Xote: Recent 

 changes to standardize tiie data processing pro- 

 cedures at the C(i Oceanograpliic Unit have re- 

 sulted in some program name changes as well as 

 minor changes in the way in wliicli future data 

 will be processed.) 



Quality Control 



STD data were quality controlled by compar- 

 ing STD analog trace and DDL values with 

 teiii[)erature and salinity values obtained from 

 Xiskin bottles attached just above the under- 

 water sensor unit. Quality control (QC) samples 

 were taken at tlie surface, 200, 500. and 1000 

 meters where possible. Tlie Xiskin bottle was 

 equipped with protected (and for the 500 and 

 1000 meter samples, unprotected) deep sea re- 

 versing tliermometers. The thermometers were 

 allowed to soak for six minutes at each QC depth 

 to reacli e(|uilibrium before the Xiskin bottle was 

 tripped. The conductivity ratios of the quality 

 control samples were detei-mined using an induc- 

 tive laboratory salinometer and were (•(mveited 

 to salinities utilizing the metiiod established in 

 the International Occanograpliic Tables pub- 

 lisiied jointly by UXESCO and the Xational 

 Institute of Oceanography of Great Britain 

 (1966). 



The ditl'eience between STD and (juality con- 

 trol values of temperature, salinity, and depth 

 were plotted against the station numbers in the - 

 order in which they were occupied. Inspection 

 of the plots indicated that the depth and tem- 

 perature values should be corrected by values 

 which did not change throughout the cruise. The 

 correction for salinity values appeared to go 

 through three phases, becoming worse as the 

 cruise progressed. The final corrections shown 

 in Table 1 were based on the average corrections 

 for surface and 1000 meters. The correction for 

 intermediate values was linearly interpolated. 

 The data for 200 anil 500 meters indicated that 



s 



