Stage 3 - Observations, background and priors¶

The solution of Equation 1 requires a prior estimate of s (called s_p or sprior). The user must create the sprior file with dimensions ntimesteps x nlandcells, so there is a value for each land cell at each time step.

Equation 1 also uses z, a vector of measurements for use in the term $(z-Hs_p)$ . This term is defined as

$(z-Hs_p)$ = (observation - bg - Hs_p - Hother )

The program make_z.py makes these calculations and creates one file used in the inversion,

zhsp.txt - contains $(z-Hs_p)$ , dimensions (nobs)

and several additional files for information:

obs-bg.txt - Observation values minus the background values.

hsprior.txt - H convolved with sprior.

Prerequisites¶

observation list - File with observation values, one value per line for each receptor.

background list - File with background values, one value per line for each receptor.

sprior - contains s_p. Can be either a numpy save file, dimensions (ntimesteps x ncells), or a netcdf file, with a variable named ‘sprior’ and dimensions (ntimesteps x nlat x nlon)

H blocks created by hsplit.py.

(Optional) hother, one value per line for each receptor, with additional values to subtract from z.

Usage¶

make_z.py is run with:

Usage: make_z.py [--other=otherdata] [-c configfile] obsfile spriorfile backgroundfile

where

  Reguired arguments
        obsfile            : file with observation value, one line per receptor.
        spriorfile         : file with sprior data. It can be either .npy or netcdf format.
        backgroundfile     : file with background value, one line per receptor. Default is 'bg.txt'.

  Optional arguments
        -c configfile      : file with configuration data.  Default is 'config.ini'.
        --other=otherdata  : file with other values to subtract from observations, one line per receptor.


  Output:
        zhsp.txt    - Observations minus background minus hsprior minus hother
        hsprior.txt - H convolved with sprior
        obs-bg.txt  - Observations minus background