PTEND (Pressure Tendency)
                        ----- --------- ---------

Parts of the full reanalysis archive contain the pressure tendency (PTEND).  
The term was dropped for the following reasons.

PTEND is dP(surface)/dt at the analysis time and has NOT been averaged over
a time interval.  The average meteorologist considers the pressure tendency
to be the pressure change over perhaps a 1 or 6 hour interval.  On the other
hand, a physicist measuring the instantaneous dP/dt would measure the effects 
of sound and gravity waves as well as the 6-hour trend.  Since PTEND was
based on dPsfc/dt at the analysis time, it was influenced by high frequency
noise like gravity waves (from the initialization?).  PTEND was simply not 
what meteorologists expect.

In a sample month, the average PTEND was ~30 mb/day in some places.  Obviously
dP/dt at t=0 is not a great measure of dP/dt averaged over a finite amount of 
time.  The instantaneous values of dP/dt can be quite large especially in the 
small scales which may be slightly out of "balance".  There has been the
suggestion that PTEND was not calculated optimally.  (Since PTEND is not
used by the model, it has not undergone much examination.)