temperature, and depth as well as the total 

 number of values retained are printed. This 

 printing gives a check that determines whether 

 any unrealistic values were used in the subse- 

 quent filtering and averaging routines. 



The third subroutine, SMOOTH, performs 

 low-pass filtering to the three channels and deci- 

 mates. The low-pass filters are a running mean; 

 the numbers for the averages depend on the re- 

 cording and drop rates. Since our procedures 

 normally produce about five values every meter, 

 a running mean of 5 is applied. Decimation 

 occurs after the second running mean and depth 

 values are rounded to the closest 1-m. interval. 



The final subroutine, INTERP (interpret), 

 is an averaging and interpolation package. Values 

 of temperature and salinity that occur in the 

 same 1-m. interval are averaged. Where there 

 are no values in a meter interval, one is linearly 

 interpolated from adjacent values. After this 

 procedure, subroutine SALTFX (salt fix) is 

 called and salinity values are corrected for swift 

 changes in the temperature gradient according 

 to the formula: 



(IS) RiKS ' 



where S' is the apparent salinity 



?- is the rate of change of temperature 



dz with depth 



K is a constant (- -0.09 p.p.t. per i°C.) 



t is a thermometer time constant 

 e0.35 sec.) 



and R is the drop rate. 



A final output statement follows this last sub- 

 routine. 



Figure 3 represents the final output from 

 the data-processing program. Preceding the data 

 are the station number and the total number of 

 observations within the bounds set in TRANS 

 and transmitted to subroutine SMOOTH. Below 

 are the maximum and minimum values of 

 salinity, temperature, and depth used by the 



smoothing subroutine. Finally, the number of 

 data sets transmitted from SMOOTH to the final 

 subroutine INTERP are given. The data interpo- 

 lated to a 1-m. interval follow below. 



A comparison of the DDL output with the 

 analog output of the STD is presented in fig- 

 ures 4-6. A 600-m. station was chosen for the 

 comparison as it presents most of the features 

 normally encountered on an STD cast. Figure 4 

 is a reproduction of a cast made near the equa- 

 tor in the eastern Pacific. The surface tempera- 

 ture and salinity noted at the top of the trace 

 were determined from a continuous recording 

 surface TS recorder, periodically checked by 

 bucket temperature and surface-water sample 

 salinity. The numbers adjacent to the profiles 

 represent the salinity scale (4) and the tempera- 

 ture scales (6, , 3) used during the cast. The 



salinity trace is displaced upward (toward a 

 shallower depth), by 5 m. on the depth scale, 

 from the temperature to allow the two pens 

 to cross without interfering with one another. 

 This particular paper does not have the scales 

 printed directly, but they are identical to those 

 at the bottom of figure 5. The spikes in the 

 salinity trace are a feature common to almost 

 all STD casts. They are considered to be a re- 

 sult of a failure of the electronic system in the 

 salinity sensor to respond to sudden changes of 

 the temperature gradient and do not reflect the 

 true salinity at those locations. 



Figure 5 is a computer generated plot of 

 the 1-m. values as output from the data pro- 

 cessing program. The scales are identical to 

 those of figure 4. The temperature profile is 

 identical to the analog plot reproducing fully 

 temperature inversions and sudden changes in 

 gradient. The salinity trace, when the vertical 

 pen displacement is accounted for, is also re- 

 produced with some spikes totally eliminated 

 and others partially eliminated. The failure to 

 eliminate all the salinity spikes reflects the fact 

 that not all the spike-forming processes are 

 known. The manufacturer has recently noted 

 that there is an intermittency in a portion of 

 the electronic system that responds to a sudden 

 change in temperature gradient as well as a lag 

 response function in the operation of the con- 

 ductivity to salinity circuitry. The manufacturer 



