Changes between Version 3 and Version 4 of ticket/1299
- Timestamp:
- 2014-10-29T11:46:13+01:00 (9 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
ticket/1299
v3 v4 15 15 Implement an on-line vertical interpolation of BDY inputs needed to allow easy switching between vertical coordinate systems without the need to alter input boundary files. 16 16 17 Files affected 18 : 17 Files affected: 18 19 19 * fldread.F90 20 20 * bdydta.F90 … … 24 24 * iom.F90 25 25 26 Testing carried out on the AMM12 configuration in the trunk code. Six experiments were carried out to check the consistency of the code. These were: 26 27 28 # Constant T/S boundary conditions on original vertical grid, with nb_jpk_bdy = 51 (first sanity test) 29 # Constant T/S boundary conditions on staggered vertical grid, with nb_jpk_bdy = 51 (gdept(jpk = [1 51] ) same as native grid all depths in-between offset from native grid, second sanity test) 30 # Realistic boundary conditions on original vertical grid, with nb_jpk_bdy = -1 (i.e. should be the identical to trunk code) 31 # Realistic boundary conditions on original vertical grid, with nb_jpk_bdy = 51 (i.e. should be the identical to trunk code within rounding errors in interpolation routine) 32 # Realistic boundary conditions on staggered vertical grid, with nb_jpk_bdy = 51 (gdept(jpk = [1 51] ) same as native grid all depths in-between offset from native grid) 33 # Realistic boundary conditions on a reduced z-level staggered vertical grid, with nb_jpk_bdy = 31 34 35 The datasets for the offset and reduced level input data where generated using the interp routine in matlab. 36 37 Results from the tests: 38 39 Outstanding issues: 40 41 How best to handle the baroclinic velocities in the case of a bathymetry mismatch between input data and native grid: 42 # Set velocities to zero in depths on native grid deeper than bathymetry in input file 43 # Down fill velocities for missing data 44 45 in all cases a barotropic correction has to applied. At present a conservative remapping option does not exist. This should be a straight forward addition of a subroutine in fldread.F90 27 46 28 47 ----