The following are proposed corrections as of 28-May-04. It is
expected that all of these will be implemented, with the date of
implementation dependent upon the completion of tasks by the Software
Engineers (note that the SEs are over-committed, given upcoming field
projectes and other tasks). Text is taken from an e-mail to the
system SEs.
- Ka and S-Pol beams do not match in azimuth. The direction of
mis-match changes with scanning direction. This is likely due
to S-band beams being timed at the start of the beam, whereas
Ka are timed at the end of a beam. (Note that the discussion
on this is ongoing -- Frank Pratte thought the difference was
the other way!)
We need a routine that will time-shift the Ka data by one beam, and
match it to the S-band beam.
- S-pol gate data do not match Ka gate data. This is obvious from
hard target studies. The S-band data needs to be moved outward by
one (integer) gate. This includes a lot of parameters!
- The Ka band HH noise power is in error in the Ka housekeeping, and
has an impact on calculated values of Ka power (P_HH_K), which goes
into Ka reflectivity (Z_HH_K). Housekeeping shows noise power of
-110.0 and -110.5. Frank P's best guess value is -112.1. This
makes a real difference at low rec'd power, and should be fixed.
After consulting with Frank, we may also want the option of putting
in an artifically *low* value of noise power. Something selectable
like -200 dBm might be nice. This will effectively nulify the
noise correction to the power measurement (or maybe we could just
turn off the noise correction).
Note that the cross-polar Ka noise power is also in error.
- We need to adjust for a bias in S-band Zdr. Analysis indicates
that we have to add .08 dB to all Zdr values.
- We need to correct the Ka copolar radar constant for the first
portion of the experiment. The correction can be applied to Z_HH_K
(similar to the Zdr correction at S-band), or we could attack
things up front by changing the radar constant and using the new
value to re-calculate Z_HH_K fgrep P_HH_K and range.
- The hardest job is an adjustment of Z_HH_K based upon an expected
P_HH_K test pulse power, combined with an automatic analysis of a
sweep-by-sweep variation from that nominal value.
For this, you need to have a declared nominal TP power
value, compute the average TP power for a sweep (being sure to
filter out any meteorological echoes!!!), then adjust all the
echo powers by the amount of the difference, do the noise
correction, and compute the reflectivity.