SMOBA
           Stratospheric Monitoring Ozone Blended Analysis

                                 by
            Shuntai Zhou, Shi-Keng Yang and Alvin J. Miller

                Climate Prediction Center/NCEP/NWS/NOAA


Abstract:

        A 3-D  daily SBUV/2 ozone analysis with TOVS ozone for polar nights

        Input to SMOBA: NESDIS SBUV/2 from NOAA satellites (19,16,...), 
        layer ozone amount and pressure level ozone mixing ratio from 0.1 hPa to 1013 hPa. 
        Total column ozone also included.

        Analysis method:Successive Correction Method
                        Fill in polar nights with TOVS  ozone with profile
                        fixed from the boundary.
                        In case of missing TOVS, use interpolation.

        Output from SMOBA:24 levels O3 mixing ratio from
                        1000 hPa to 0.2 hPa, and total column ozone
                        at 2.5 x 2.5 deg.

        10 day rotating files in FTP site: ftpprd.ncep.noaa.gov

                        Directory: /home/people/cpc/wd53sz/SMOBA


==============================================================================

(1) The raw data (retrieved by NESDIS)

        Ozone mixing ratio (ppm) and total column ozone (DU) are
analyzed from daily SBUV/2 (currently on the operational NOAA satellite 19)
measurements. The retrieved ozone data include 15 levels of mixing ratio.
Vertical resolution is very coarse in the troposphere (only 5 layers).
The data are arranged along satellite orbit tracks in the BUFR (Binary
Universal Form Representation) format.

(2) The analysis

        The objective analysis is performed from the orbital BUFR data
to a global, 2.5 x 2.5 degree latitude-longitude grid by a commonly used
successive correction method (SCM). The output includes mixing ratio
for 24 levels ranges from 0.2 hPa to 1000 hPa (see data format below).
Note that due to the different vertical resolution of the
raw data, the reliablity of ozone mixing ratio is best above 30 hPa, good
between 100 hPa and 30 hPa, and poor below 100 hPa. The total ozone is
directly from the BUFR data, which is derived by NESDIS using retrieval
and climatological information. No vertical integration is done in the
analysis scheme.

(3) The polar night gaps

        Because SBUV/2 is a sun-lit instrument, there is no data
in polar night. To fill in the polar night gap, we use TOVS total ozone
as auxiliary information. The shape of polar ozone profile is set to be the
same as that of SBUV/2 ozone at the boundary of polar night, and the polar
total column ozone equals TOVS measurement. Since TOVS ozone is slightly
different from SBUV/2 ozone (~5% at the boundary of polar night), a linear
adjustment is applied to the polar ozone so that the filled data joint at
the boundary smoothly. If TOVS ozone is not available for any reason,
we use interpolation to fill in the data gap. Therefore, there are no gaps
in the data analysis files.

(4) Total ozone comparison with other measurements

         Compared with other total ozone measurements such as TOMS and
TOVS, the general differences are within 5-10%, depending on time and space.
In the tropics there is little difference (<5%) from TOMS, but there are
larger difference from TOVS (>10%). Those differences are partly due to 
different observing time and partly due to calibrations. 

(5) The analysis data format

  FILE NAME:   ozyymmdd.dat    (yy:year, mm:month, dd:day)
  DATA FORMAT: ascii array A(144,73,25)
             i=1,144 ------ longitude 0 to 357.5 (2.5 degree grid)
             j=1,73  ------ latitude 90N to 90S  (2.5 degree grid)
             k=1,25  ------ 0.2 0.3 0.4 0.5 0.7 1.0 1.5 2.0 3.0 4.0
                            5.0 7.0 10. 15. 20. 30. 40. 50. 70. 100
                            150 200 250 300 hPa mixing ratio
                            k=25 for total column ozone
                            ozone mixing ratio below 300 hPa is set
                            to be uniform
             (((A(i,j,k),i=1,144),j=1,73),k=1,25)  (1x,8f9.3)