them, file by file, and lists the binary length of 

 the first record as well as the total number of 

 records per file. Any parity errors in the records 

 examined are also listed. 



The ideal field tape contains no parity er- 

 rors. The first record of the first file is an infor- 

 mation record and, in the format used by us, is 

 a three-digit number signifying the cast number 

 for that particular cruise. The binary length of 

 this record is always 1. The second file consists 

 of the data recorded by the data logger and may 

 contain any number of records depending, among 

 other things, on the maximum depth attained 

 and the drop rate. A 500-m. cast at a drop rate 

 of 60 m. per minute has about 50 records per file. 



In its present form the data logger is de- 

 signed to produce a binary data record length of 

 52.3 xhus, in the ideal field tape the files alter- 

 nate between an information file containing only 

 one record with a binary length of 1 and a data 

 file with many records, all with a binary length 

 of 52. The preliminary listing of the field tapes 

 provides the programmer with a picture of how 

 far his tapes deviate from the ideal. If the con- 

 tents of the first record are printed during the 

 preliminary listing, the cast numbers may be 

 identified with individual data files. 



The next step in the data processing is to 

 produce, from the field tapes, a high-density tape 

 which is free of parity errors and other anomalies 

 which confuse the tape translation. In the trans- 

 fer from low- to high-density tape, we have cho- 

 sen to eliminate all records containing parity 

 errors and records not of binary length 1 or 52. 

 Our experience is that we lose no more than 

 3 percent of the original data in this way. The 

 high-density tape, free of tape errors, is then 

 considered to be the basic data; the original 

 field tapes are erased, checked and readied for 

 the next cruise. 



THE PROGRAM 



The program and subroutine functions 

 are outlined in figure 1, and a listing of the 

 program, as run on the CDC 3600 at UCSD 

 (University of California, San Diego), is pro- 

 vided in the Appendix. The main program 



RDEDTP (read and edit tape) reads in, from 

 cards, a list of the files to be translated from 

 the basic tape and a list of station numbers that 

 are to be associated with data lists. The file 

 containing the station number is read and trans- 

 lated if it is in the proper format. If it is missing 

 or in an improper format, as determined from 

 the preliminary tape listing, the proper station 

 number is determined from a logbook for the 

 data logger and is read in from a card. 



The first subroutine, TRANS, translates, 

 record by record, the digitized frequencies 

 into salinity in parts per thousand, temperature 

 in degrees Celsius, and depth in meters. Maxi- 

 mum and minimum bounds are specified for 

 the depth so that any values outside these 

 limits are rejected. 



During the field operations, the sensor 

 package is sometimes temporarily stopped at an 

 intermediate depth to make adjustments to the 

 pens or the winch. The data logger is usually 

 left running on these stops, but the records 

 are of no use in producing a vertical profile; 

 subroutine BASKET is accordingly called to 

 delete the records where the depth has not in- 

 creased. The final control is a counter which 

 provides for a jump out of TRANS before 

 memory overflows can occur. 



A sample output of the record produced by 

 subroutine TRANS is given in figure 2. The 

 fourth column of each set is an absolute counter 

 for the file, which increases only when an accept- 

 able set of salinity, temperature, and depth is 

 translated. When the lower depth bounds are 

 exceeded, the counter is not increased and that 

 set of data is not saved for transmittal to the 

 smoothing subroutine. The 0's printed at the 

 beginning of the file in the first record indicate 

 nonacceptable data, since the depth is less than 

 the minimum of -0.2 m. set in the subroutine. 

 In the first record of this file only 16 sets of 

 salinity, temperature, and depth were acceptable. 



Between TRANS and the next subroutine, 

 the maximum and minimum values of salinity, 



^The magnetic tape record is 416 tape frames long, 

 which is the equivalent of 52 computer records. 



