A software error in the Ka-band radar processor code created a
condition whereby returned power from a given radar pulse could be
mis-directed into the wrong location of the Ka beam circular buffer.
On certain specific days, pulse energy could be added to a beam that
was 20-beams off of the current beam being computed. Perhaps one or
two erroneous pulses would be added to the wrong beam. One pulse out
of the 50 desired pulses equates to an echo that is suppressed by 17 dB
( 10*log(1/50) ~ -17 ); actual ghost echoes appear to be
closer in energy to the original echo than the 17 dB suppression would
suggest.
The net effect is the creation of "ghost" echoes that are 20-beams (about 16°) out of sync with the real echo. Such a problem has been noted for 30-Dec, and 7,11, and 21 Jan (note that December has not been fully inspected).
The appended figure clearly shows ghost echoes for 7-Jan, with the image shifting back and forth between S-band and Ka-band reflectivities.
Additional information is available:
[Do shift-Reload to re-animate this image.]
