Modifications to STILT - Approach similar to Lin and Gerbig ’05
- Uncertainties comparable to uncertainties due to wind errors
- Paper in prep.
Modifications to STILT - input: configuration file “namelist”
- Grid specifications (projection, limits)
- Time frame
- output:
- EVI, LSWI & Vegetation map at desired resolution for one year (min)
- netcdf format
- Uses: Modis reflectances MOD09A1 from local database
- Tool tells what tiles are required for downloading
Modifications to STILT Coupling hourly emission fluxes within Trajecvprm() EDGAR emission converter - get.edgar() to create hourly fluxes at desired resolution (needs upgrade to create ncdf files)
coupling to TM3 global model as Boundary Condition - TM3 runs using fluxes from inversions, coupled to Biome BGC
- Check if nesting improves in forward runs
Nested inversions coupling STILT + TM3 adjoint - European nest for high resolution (~ 20 km) (Kristina Trusilova)
- Tall tower data plus global network as input
- At specific tall tower locations in Europe + aircraft data CERES
Future: keeping STILT stable - Method: set of hymodelc test runs with a result control facility
How (suggestion): - @ Harvard Grid computer
- each group provides a (set of) test runs with all required inputs
- SK provides a fuzzy checker tool to compare PARTICLE.DATs as core of semi-automated verification (random off, can only check when results are not supposed to change)
- developer should run one batch file only which is part of the STILT CVS repository itself
- it takes f90 files from user (scp to local machine) and compiles, runs and compares on GRID
Dostları ilə paylaş: |