From: OKA Eitarou [mailto:eoka@ori.u-tokyo.ac.jp] 
Sent: Wednesday, August 16, 2006 2:15 PM
Subject: North Pacific Argo float data: An Important Notice 2


Dear "North Pacific Argo float data set" users,

Yesterday, I updated my dataset, adding the data of July 2006.
http://ocg.ori.u-tokyo.ac.jp/member/eoka/data/NPargodata/
This month, I made a SMALL CHANGE in the DATA FORMAT.

Specifically, I changed the header of each profile.

  header example: " 151 JM 5900649  66 20060703  23.530 210.258 115    "

As explained in README, the header structure is as below.

   1-4 th characters :   sequence number         (151 in the above example)
      .....
   38-44th characters : longitude                      (210.258)
   45th character :       (space)
   46-48th characters : sampling layer number   (115)
   49th character :       (space)
   50-52th characters : QC information              ("   ")

This time, 45th character (space) and 46-48th ones (sampling layer
number) were unified, and changed as below.

   1-4 th characters :   sequence number 
      .....
   38-44th characters : longitude
   45-48th characters : sampling layer number
   49th character :       (space)
   50-52th characters : QC information

The digits of the sampling layer number increased from 3 to 4, 
but the space on its left was eliminated, so the other parts of
the header are unchanged.  The new 4-digit sampling layer number
is right-aligned as the old 3-digit one, so if the sampling layer 
number is less than 1000, the header looks exactly the same as before.

In my dataset, there haven't been any profile with sampling layer number
more than 1000 so far.  However, recently the profiles from some floats
using the Iridium system have large sampling layer numbers around 500.
In a near future, some profiles might have sampling layer number
more than 1000, and that's why I made the format change this time.

Please ask me if you have any question.

thank you,
Eitarou

--------------------------------------------------------------------