We are supplying the following files which list position and
 atmospheric conditions at ARGOS drifters located around the
 SHEBA ship.  All files are ASCII.  All files are processed,
 files which have been despiked and compared to other buoys' files
 to attempt to eliminate unreasonable values.  Although we usually
 use an Akima cubic spline to fill short gaps, we have not done so 
 with these files, to allow people to spline or fill as they see fit.
 Instead, gaps in data are marked 1.e35.  All the  files except two
 contain hourly-sampled data, but all files have frequent short gaps
 (several hours) when data were not received.   Two files list only
 ARGOS times and positions; these files contain unevenly-spaced data. 
 
 There are several types of files:
1) Files containing time (yeardayhr) GPS position, air temperature, and 
 snow temperature, one point per hour.  The temperature sensor range 
 bottomed out at -40C,  so to avoid the appearance of constant cold 
 temperatures we labelled -40C temperatures as missing values (1.e35).  
 Few buoys were around in  the summer, so we couldn't really intercompare 
 the summer temperatures,  which may be high because of insolation and snow 
 melt (around the snow thermistors).
 
 These GPS+temp files are:
 07100.gpt  This buoy didn't like the cold and only started sending
 reasonable amounts of data on 11 May 1998 (1998131).  It had a gap
 in temperature from 199814020-199814108 and from 1998165 to the end
 of data on 1998184.

 20726.gpt  This buoy was deployed on 199727523 and crushed on 199801423.
 It had no major gaps during that time.

 22204.gpt This buoy was deployed on 199727802 and transmitted until
 199825520.  It had fairly frequent ~12 hour gaps in the winter (1997358
 to 1998057), and its snow temperature was questionable after 1998131, 
 although they are included to day 1998163.

 22206.gpt This buoy was deployed on 199727600 and was crushed on 
 199732008.  Its temperature data had frequent short gaps, although
 the position data was fairly good except for gaps around 1997295.


 2) GPS-only buoys
 07105.gps   This buoy only measured GPS position.  It was deployed on 
 199728004 and transmitted though July 1999, although only the data 
 through 199828304 are presented here.  It had two gaps in GPS data,
 one from 1997348 to 1997355 and the other from 1998273 to 1998275.


 3) ARGOS-only buoys
     We report only the ARGOS positions from two buoys.
 The format of these files is time(YearDayHrMn), latitude, longitude,
 and location class.  Unlike GPS positions where the buoy actively
 determines its position at a given time, ARGOS positions are determined
 by the satellite.  These files therefore contain unevenly spaced data,
 which is why we specify the minutes in the time word.  The location
 class, determined by the satellite, specifies how accurate the location
 is.  We only present positions with location classes of 3 (estimated
 accuracy better than 150m) and 2 (150m < accuracy < 350m).  

 02159.apos This file starts on 1997287 and ends on 1998260.  There is a 
 44-day gap in April-May (from 1998094 to 1998138).

 22205.apos This file lasts from 1997281 to 1998261 with relatively few
 gaps; the longest is a 7-day gap in August 1998.

 4) ARGOS position + temperature and pressure
  These buoys had a transmission "duty cycle" to lengthen battery life.
 Therefore, positions are relatively sparse and unevenly sampled, while
 temperature and pressure, which are measured every hour and stored for 
 transmission are well reported. The files list the time (YearDayHrMn)
 and position if that position was determined within  30 minutes of the 
 temperature and pressure.  Remaining columns are time temperature and
 pressure were measured, and the temperature  and pressure.
 
 23676.apt lasts from 199729223 to 199814618.
 26696.apt lasts from 199729221 to 199821113.