Opened 18 years ago
Closed 18 years ago
#90 closed defect (wontfix)
Refractivity below surface
| Reported by: | frae | Owned by: | frcm |
|---|---|---|---|
| Priority: | normal | Milestone: | 1.0 |
| Component: | ropp_fm | Version: | 0.9 |
| Keywords: | Cc: | dave.offiler@…, michael.rennie@… |
Description
When running ropp_fm/tools/ecmwf2ro_1d on a netCDF file that has a fixed e.g. 200m geopotential height levels resolution starting from the ground (given to the FM in geop_refrac) I get rather funny refractivity values when I do not provide any background data below let's say 1000m geopotential height since the surface is at 1000m. The FM seems to extrapolate the values below 1000m from the ones above 1000m. This is a behaviour I wasn't expecting, one might want to use a fixed geopotential altitude grid which sometimes is below the ground, in particular when combining several occultations in one netCDF file.
Here is a netCDF input file to ecmwf2ro_1d that shows this behaviour:
ftp://sat.physik.uni-bremen.de/incoming/axel/fm_example.nc
Am I missing something here? Is some other input required?
The forward model does indeed extrapolate; it is the responsibility of the user to ensure that the results are used in a meaningful way. In particular, the user should make sure that output levels given to the forward model are not unphysical (e.g., below the Earth's surface). This is a performance issue: forward models may be called a large number of times, and it is simply not feasable to carry out consistency checks at ech iteration without performance penalties.