A rather crude, yet effective, relational database was created to handle the newly collected data of the cuise, along with the available historical data, from the MEDATLAS database [11].
Once collected, the XBT (or XCTD) data were just edited for the headers and for very bad and suspicious data and immediately put into the database. The depth readings were corrected according to the UNESCO recommendations [34]. A few procedures were then prepared for the browsing and plotting of the data, along with the procedures for the creation of the velocity profiles for the depth processing. These last accounted primarily for the data in the upper layer down to 700-800 m. After this depth the profiles were integrated with the best available data found in the historical databases or within the ISDGM profiles. The analysis led to the subdivision of the surveyed area into some major areas with the same hydrologic properties, and the relevant transition zone between them.
At this point, the data were used to provide the best sound velocity profile for any grouping of lines that crossed the geographical blocks of approximately 20x30 primes (or more) in latitude and longitude, respectively, that divided the survey area for processing. This sound velocity profile was then used for the depth processing phase in NEPTUNE (as it will for CARAIBES). Some software routines were built for the handling of data and for the automated production of the sound velocity profiles, according to the formats required by the KS EM12 BDU and software (AML and NEPTUNE).