January 15, 2009 Meeting Summary |
Vijay Tallapragada presented a status update on HWRF T&E for the 2009
implementation. The finalized changes to the HWRF for the 2009
pre-implementation testing include: initialization-related source code
edits, gravity wave drag parameterization, a bug fix for radiative
fluxes in the nested domain, land-surface temperature bug fix in the
nested domain, HWRFPOST will include additional variables, faster
copygb, a bug fix in the read/write auxiliary files, a tpcpost script
that will exclude the generation of swaths where the storm dies within
the first 6 hours, and an inclusion of POM output in the archives. Vijay
also mentioned a unification of HWRF source code so it may be used by
all developers. The unified code has features such as uniform code for
both HYCOM and POM, inclusion of gravity wave drag parameterization, sea
spray parameterization (currently turned off), land-surface temperature
bug fix, inner domain short wave radiation bug fix, and a restart/high
resolution option (currently turned off). The framework for the 3-way
coupling of HWRF-HYCOM-WaveWatchIII has also been added to the scripts
for use in the future. As requested by NCO, more error checks and alerts
were included in the scripting to aid in diagnostics. Vijay anticipated
that the system will be ready for testing by January 22. Testing will be
performed on 2008 Atlantic and East Pacific storms using the new GFS/GSI
Q1FY09 runs. It is also expected that the transition to P6 will be
completed by the end of January for the HWRF, so testing can begin on P6.
Vijay mentioned that the HWRF might fail if it could not couple to POM
or HYCOM properly. The restart option would be helpful in this case, so
turning that option on in the model was mentioned. While the restart
files could take up a lot of disk space, only saving the restart files
from the most recent runs before writing over them in subsequent cycles
would solve that problem.
Steve Lord brought up the topic of using regridding instead of
relocation in the HWRF. Regridding would involve moving the whole grid
so the hurricane was in the correct place. Since grid stretching might
occur, configuring the grid so the stretched areas are further from the
storm would be necessary. More discussion on this topic will follow.