PROCEDURES 



Oceanographic Sampling 



At I'iuh station an STI) (Salinity-TempiMa- 

 tnro-l)pi)tli) cast was taken to near bottom or to 

 a depth of HOC) meters. Tlie data was collected 

 on a Plessey Environmental Systems Moilel !»040 

 S/T/D Enviionmental I'rofilinjr System (STD) 

 (serial number r)3i:}). The data were recorded 

 on an analojr trace and also di-iitally on ma<rnetic 

 tape. The dijxital recordin<r was made by a 

 Sonycraft Di<rital Data Lojrger (DDL) manu- 

 factured tmder ('oast (luard contract CG-12, 

 778-A. Four channels of informati(jn were 

 sampled at rates of 0.5 or 1.0 scans per second. 

 STD freciuencies represent in<r depth, tempera- 

 ture, and salinity were converted to binary coded 

 decimal (BCD) and recorded on a 7 channel 

 IBM compatible ma-inetic tape at a bit density 

 of 200 bpi. The resolution of the DDL system 

 is ±one hertz. One liertz corresponds to 

 .0O344°/„„, .018°C. and l.!)0 meters in salinity, 

 temperature, and depth respectively. The fourth 

 information channel was available for recording 

 sound velocity on the DDL, i)ut it was not used. 

 The tape format for each STD cast consisted of 

 3 sets of station data such as station number, 

 position, date, and time followed by any number 

 of data records, dependinjj on the maxinnim 

 depth and lowerinjr rate of the cast. Each record 

 consisted of the tempeiature and salinity infor- 

 mation at 100 depth levels. Thus, an average 

 one thousand meter cast was composed of about 

 1200 data levels lecorded on approximately 1'2 

 records. Five computer programs were devel- 

 oped by CG OCEANOU to reduce t lie number 

 of data levels to a more numageable figure of .".() 

 to 100 data levels at standard depths and inflec- 

 tion points which would still accurately represent 

 the original water column. 



The computer programs were developed for a 

 Control Data Corpf)ration (CDC) 3300 com- 

 1. liter. A flow diagram of the proce.ssing proce- 

 dure des(rii)ed lielow is shown in figure -1. Tiie 

 first program. NE\\'I)L. input tiie (m deck depth 

 frequency <'f tlie deptli sensor, and read the 



records to be processed from the magnetic tape. 

 The digitized fre(piencies were translated from 

 BCD to engineering units of depth (meters to 

 tenths), temperature (C° to hundredths), and 

 salinity (°/„„ to hundredths). The values were 

 printed out so that an initial check of the data 

 c.uld be made. In addition, a tape output 

 (NKWDL tape) was written as an input to the 

 next program. "With a rapid sample rate such 

 as 0..") second, a specific depth level might show 

 up several times. While tiiese temperature and 

 salinity values weie always close, they generally 

 did not agree exactly, probably as a result of 

 sensor lag. The output from the first program 

 was normally aroimd 1200 levels of data for a 

 1000 meter cast. 



Program AVCOK averaged data levels in- 

 putetrfrom the NEWDL tape at the same depth 

 level. AVCOK accepted seciuential levels until 

 a deeper level was reached; then it began the av- 

 eraging for the next level. Therefore if, due to 

 the^shri) rolling, tlie STI) dips to a lower level 

 and tlien ret\irns to the original level, the data 

 at the original level subsequent to the roll will 

 not be included in the average. During the 

 AA'COK processing.' corrections are made to 

 temperature and salinity as discussed in the fol- 

 lowing section. Tlie output of AVCOR is a 

 printout and a magnetic tape (AVCOK tape). 

 Tlie i)rintout of temperature, salinity, and com- 

 puted sigma-t was quality controlled by remov- 

 ing samples which caused averaged sigma-t 

 vabies to decrease more than 0.2, 0.05. or 0.02 per 

 averaged data level within 0-100 meters. 100-300 

 meters, and deeper than 300 meters respectively. 

 Use of these criteria occasionally permitted 

 data to pa.ss which indicated large instabilities 

 in the water column, as revealed by computation 

 of the stability or E value (Sverdrup. et al. 1942, 

 pp. 416-418). This usually occurred only over 

 small intervals. (Although such data might be 

 (piestitmed, tiie data has not been rejected; this 

 will permit other investigators to .haw their own 

 conclusions as to whether or not to use the data. 



