The following text describes the contents of the SAMBA subsurface float files postfixed by .fin* or .raw ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ALL DATA are from The SAMBA experiment (a component of the WOCE subsurface float program) All floats were MARVOR (Tekelec company) multicycle sub surface floats, with one exception (the v19 Seascan company float). The data files give the float positions during their deep journeys. Actually the floats were used as water parcel followers near 800 m depth for 2-month or 3-month periods (depending on their scheduled mission), with 2-day or 3-day periods (respectively) at the sea surface , between the deep cycles. These short surface drifts were necessary to communicate with system Argo satellites so that data acquired at depth can be transmitted for later analysis. The files postfixed by .fin* (.fin or .finbis or .finman or .finmanbis, ...) give the deep float positions during their deep journeys, whereas those postfixed by .raw give only the first and last positions (that is the last position while the float was at the surface before plunging and the first position after surfacing 2 or 3 months later). This is because the floats were positioned acoustically once per day, and for a few floats and cycles it was not possible to retrieve any correct sound signal to estimate deep positions. Surface positions were given by system Argos. There is one file for each cycle (a cycle comprises the float descent to the depth of 800 m (or rather 800 dbar), the free drift of the float near that depth, the ascent to the surface and finally the surface dift proper. Data format is the FLOATER format designed by Phil Richardson (WHOI): (3A2,1X,I4,4I2,F7.3,F8.3,5X,I1,2X,F8.3,2X,F8.3,3X,F7.2,3X,F7.2,2X,F8.0) with the following parameters: FLOAT ID Position 001-006 DATE (year, month, day, hour, min) Position 008-019 Be careful however, since before the year 2000, the year is generally given on 2 characters, thus the Date on positions 010-019 (for example 1999 is given as 99). LATITUDE Position 020-026 LONGITUDE Position 027-034 VELOCITY Position 040 This information indicates the algorithm used to estimate the velocity components.Actually, all float velocities have been estimated with cubic splines fitted to Lanczos filtered (with a 3-days cut-off period) latitude and longitude time series. default is 0, otherwise 1. VEAST Position 043-050 (cm/s) default is -999. VNORTH Position 053-060 (cm/s) default is -999. TEMPERATURE Position 064-070 (°C) default is -999. or -99. or 99. PRESSURE Position 074-080 (dbar) default is -999. or -99. JULIAN DAY Position 083-090 Be careful however, since here Julian days are considered to begin at 0h UT, although exact Julian days should begin at 12h UT. We have done that to get integer Julian day numbers corresponding to Gregorian days. References:Ollitrault et al. (1995) The SAMBA experiment, vol 1 SAMBA1 LAGRANGIAN and CTD DATA. IFREMER Ed., Repères Océan n° 12, 488 p.