Time: 02:00~03:30 pm, Oct 26,
2005
Participants: Dr. Huang, Dr.
Guo, Shui-Yong Fan
Shui-Yong
showed the verification diagrams of EXP04
and EXP05 on the meeting. (For saving time, the following experiments will run
only in Domain1 and so only the verification results of Domain1 will be
compared.) The scores of EXP02 were still best in all of experiments. The
scores of EXP04 were worst that indicated the significance of observation
analysis. The scores of EXP05 were not the same as EXP00 maybe because of the
different assimilation of surface observations. So Dr. Huang suggested another
experiment should be done with the same assimilation of surface observations as
EXP00. And he also thought that with the same domain configuration of BES as
Var, the options of len_scaling and var_scaling should not be tuned too much.
So he suggested another experiment should be done with len_scaling=1 and
var_scaling=1 and then maybe run an experiment with Cressman analysis as a
bench-mark.
EXP04: no 3dvar analysis and IC
(initial condition) from T213 firstguess field. This is a bench-mark that shows
how much 3dvar impacts on forecasting.
EXP05: ncar3dvar
with BMB BES and namelist.3dvar is the same as that of BMB OPR(operational
system) except the two options (chech_rh=2 and sfc_assi=2 but both are 1 in OPR).
Shui-Yong
completed the code of gen_be for mm5 and has done some initial tests, including
compiling/running from Stag0 to Stag4 and wind difference fields test. But more
tests should be done, for example, T/RH/Psfc(surface pressure) difference
fields test and transform test.
Next meeting will be at 2:00
pm, Nov 10, after Dr. Guo back from China.