TSMP-PDAF with PDAF-OMI

PDAF-OMI for multivariate data assimiliaton. It is design to handle different observation types (currently soil moisture and TWS) automatically. Additional to current observation files, the observation type has to be included (SM, GRACE).

See also create observation script for details: https://icg4geo.icg.kfa-juelich.de/ExternalRepos/tsmp-pdaf/tsmp-pdaf-observation-scripts/-/tree/main/omi_obs_data?ref_type=heads

PDAF-OMI can be switched on by the command line option use_omi.

Both global and local filters can be used. To enable multi-scale data assimilation, different localization radii for different observation types can be passed. Note that the localization radius for SM is currently in km and for GRACE in #gridcells.

If questions arise contact ewerdwalbesloh@geod.uni-bonn.de

Observation File Format

The reader read_obs_nc_type expects the following variables in each observation file:

NetCDF name	Type	Required?
dim_obs	dimension	yes
obs_clm	REAL(dim_obs)	yes
type_clm	CHARACTER(20)(dim_obs)	yes
lon	REAL(dim_obs)	yes
lat	REAL(dim_obs)	yes
layer	INTEGER(dim_obs)	yes
dr	REAL(2)	yes
obserr_clm	REAL(dim_obs)	if multierr=1
obscov_clm	REAL(dim_obs,dim_obs)	if multierr=2

Files follow the naming convention <obs_filename>.<NNNNN> (five-digit zero-padded PDAF step).

The type_clm field must be homogeneous within a file - a file containing GRACE observations cannot simultaneously contain SM or C observations.