How
do I change my old codes to read OIv2 data?
Where
can I find climatology fields?
Why
doesn't the sample code to read the data work on my system?
When
are the near real time fields generated?
Where
can I get the data in ascii?
Where
can I get the data in NetCDF?
A sample code to read the OIv2 data is available in the README file in the ftp data directory.
For users that already have their own codes written to read
the original
OI data, the most obvious changes:
- the data fields are 4-byte real rather than
2-byte integer.
- you no longer need to divide by 100.
- the fields are in individual and
yearly files
The key change that could slip by unnoticed is that the grid
is shifted
180 degrees:
OI.v1 started at 179.5W
OI.v2 starts at 0.5E
A few points on the land mask changed.
The mask file is now direct access, ieee binary, rather than ascii.
The new sample source code reads the new mask.
Data set additions:
- The relative error variance is included in a separate
field
with the weekly data.
- The estimated ice concentration is included in a separate
field in both
weekly and monthly data.
If you convert the total fields to anomalies using our
climatology,
we
strongly suggest you switch to the new climatology (see below) derived
from OIv2.
If you use the old climatology with the OI.v2 you will see large
anomalies
near the ice
edge, particularly in summer months.
A climatology derived from these fields with a base period of 1971-2000
is described at:
http://www.cpc.noaa.gov/products/predictions/30day/SSTs/sst_clim.html
and available at:
ftp://ftp.cpc.ncep.noaa.gov/wd52yx/sstclim
This is a product of the Climate
Prediction Center. Contact Yan Xue
(Yan.Xue@noaa.gov) for further
information.
You may have to swap the bytes in order to read the data on a
pc or other "little-endian" system.
There may be an option to do that easily with your fortan compiler.
Also, there is a fortran control word at the beginning and end of each
record.
This indicates the number of actual bytes of data in that record.
If your pc's compiler does not expect these, you may have to skip them.
The data file structure (for weekly data) is as follows:
(for each time step)
Header record: 1 4-byte fortran control word
8 4-byte integer words (Dates and index)
1 4-byte fortran control word
SST record: 1 4-byte
fortran control word
360*180 4-byte real words (SST values)
1 4-byte fortran control word
Error record: 1 4-byte fortran control
word
360*180 4-byte real words (relative error variance)
1 4-byte fortran control word
Ice record: 1 4-byte
fortran control word
360*180 1-byte character words (Ice)
1 4-byte fortran control word
The code can be modified to skip a dummy four byte integer
word at the
beginning and end of each read statement.
The monthly data is similar, but does not include the Error record.
The real-time weekly fields run Sunday through Saturday and are
generated early Monday morning.
We have to wait until we can interpolate weekly fields to the last
day of the month to generate a complete
monthly field. Thus, the full monthly fields are updated
according
to the following schedule:
Last
Day
of
Analysis
Month
Falls on Completed on
--------------------------------
Sunday
8th
Monday
7th
Tuesday
6th
Wednesday
5th
Thursday
11th
Friday
10th
Saturday
9th
It is best to wait until late morning (US Eastern Time) to pick up new fields to give us time to ensure the analysis ran properly.
The Data Support Section (DSS) of NCAR makes both weekly and
monthly OIv2 data available in ASCII at:
http://rda.ucar.edu/datasets/ds277.0/
If the DSS archive is not up-to-date, look for the most recent
data
at:
ftp://ftp.emc.ncep.noaa.gov/cmb/sst/oisst_v2/ASCII_UPDATE
for weekly data
and:
ftp://ftp.emc.ncep.noaa.gov/cmb/sst/oimonth_v2/ASCII_UPDATE
for monthly data
The NOAA Earth System Research Laboratory (ESRL) makes these data availabe in NetCDF at: http://www.cdc.noaa.gov/cdc/data.noaa.oisst.v2.html
Page last updated: 31 May 2012