Temperature Comparisons for GMI

GOAL:

To determine which proposed GMI assessment temperature fields are most
realistic in the lower stratosphere.

FOCUS:

The focus will be on Northern Hemisphere, middle latitude (40-50
north), 50 hPa, temperatures.  The largest fraction of the accumulated 
emissions is expected to be within this latitude band.

VALIDATION TEMPERATURES:

The proposed GMI assessment temperature fields will be compared to an
18 year record of global, gridded, NCEP/CPC temperatures.

These NCEP temperatures are interpolated to an equal area grid and
processed in 10 degree latitude bands.  The temperatures in each band
are binned into 2K temperature bins to create temperatures
distributions for each day.  The daily distributions are then 
summed into monthly distributions for all 18 corresponding months. This
temperature distribution is calculated for 3 vertical levels: 100,
50, and 30 hPa.  The final product for the validation temperatures 
consists of one 4d array: 50 (temperature bins: 180-280), 16
(latitude bins), 3 (vertical levels), 12 (months in a year).  The 50
temperature bins are further summarised as means and standard
deviations.

GMI TEMPERATURES:

Temperatures from each proposed GMI assessment temperature field will
be processed as descibed above.  Temperatures will be interpolated in
the vertical if necessary.

EVALUATION:

The grading will mainly be based on the mean temperature differences in
the 40N - 50N latitude band at 50 hPa.  The monthly average difference
will be calculated for each field and used to compute the rms difference
over the annual cycle.   The proposed GMI temperature field
closest to the climatology gets the highest grade.  Variability will
also be considered by comparing the standard  deviations for each
field. 

In addition, plans are to compare other latitudes and levels as well as
the full temperature distributions for selected months. These
comparisons will serve as general quality control checks.