Baseline actions:
1.Do some GLDAS
runs with the tile space and parameters NSIPP is currently using (JonG or
BrianC) at 2x2.5 deg.
2.Force these
runs in both near-real time and over the past 20 years or so.
RETROSPECTIVE: We currently have N.Am catchment forcing for
1979-1993 (UTexas). We need global 0.5
gridded, bias corrected forcing for 1979-near current
(Berg)?
REAL TIME: JonR is already
runing GLDAS for near-real time at 2x2.5 so that is pretty much under control (JonR). We will use the best available GLDAS
forcing (in development) for real-time mosaic runs.
3.It may be
desirable to constrain all GLDAS forcing to GPCP (even up to a month of
near-real time) to be consistent with the long-term
forcing(Urszula).
Near-term extensions:
1.It would be
quite nice to be able to assimilate
SMMR/SMMI/TRMM/AMSR data into this system to further constrain
it. JaredE and JeffW have put the Kalman Filter into Mosaic running in the GLDAS code. Need further
development and testing of the KF in the GLDAS Mosaic (Jared,Jeff,Rolf).
2.KF
observations: SMMR (79-87) is ready to
go (Owe). TRMM (98-current) derivations
are immature (JaredE) and SSMI are not existent. Push Owe to move on with SMMI(87-current), or look for them somewhere else (NCDC).
3.We will worry
about AMSR (2002-) once we see
some data.
….Then continue this development to it’s logical
glorious conclusion!
Discussion?