Fortran module for gnssro refractivity Met Office's tangent linear and adjoint. More...

Data Types
type	ufo_gnssro_refmetoffice_tlad
	Fortran derived type for gnssro trajectory. More...

Functions/Subroutines
subroutine	ufo_gnssro_refmetoffice_tlad_setup (self, vert_interp_ops, pseudo_ops, min_temp_grad)
	Set up the Met Office GNSS-RO refractivity TL/AD. More...

subroutine	ufo_gnssro_refmetoffice_tlad_settraj (self, geovals, obss)
	Set up the K-matrix for (Jacobian) for the Met Office's GNSS-RO refractivity operator. More...

subroutine	ufo_gnssro_refmetoffice_simobs_tl (self, geovals, hofx, obss)
	Given an increment to the model state, calculate an increment to the observation. More...

subroutine	ufo_gnssro_refmetoffice_simobs_ad (self, geovals, hofx, obss)
	Given an increment to the observation, find the equivalent increment to the model state. More...

subroutine	ufo_gnssro_refmetoffice_tlad_delete (self)
	Tidy up the variables that are used for passing information. More...

subroutine	jacobian_interface (nlevp, nlevq, za, zb, p, q, pseudo_ops, vert_interp_ops, min_temp_grad, nobs, zobs, Kmat)
	Calculate the K-matrix used in the TL/AD. More...

Variables
integer, parameter	max_string =800

Detailed Description

Fortran module for gnssro refractivity Met Office's tangent linear and adjoint.

Function/Subroutine Documentation

◆ jacobian_interface()

subroutine ufo_gnssro_refmetoffice_tlad_mod::jacobian_interface	(	integer, intent(in)	nlevp,
		integer, intent(in)	nlevq,
		real(kind_real), dimension(:), intent(in)	za,
		real(kind_real), dimension(:), intent(in)	zb,
		real(kind_real), dimension(:), intent(in)	p,
		real(kind_real), dimension(:), intent(in)	q,
		logical, intent(in)	pseudo_ops,
		logical, intent(in)	vert_interp_ops,
		real(kind_real), intent(in)	min_temp_grad,
		integer, intent(in)	nobs,
		real(kind_real), dimension(:), intent(in)	zobs,
		real(kind_real), dimension(:,:), intent(out)	Kmat
	)

Calculate the K-matrix used in the TL/AD.

jacobian_interface

Allocate temporary arrays
Call partial-derivatives code which calculates various quantities on model levels
Loop over each observation, doing the following
If using pseudo-levels, calculate specific humidity, pressure and temperature on intermediate levels, and interpolate these appropriately. Then calculate the refractivity gradients for each observation, interpolated from the pseudo-levels.
If not using pseudo-levels, assume that refractivity varies exponentially with height and calculate gradients.
Calculate K-matrix from the component gradients.

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Parameters

[in]	nlevp	Number of pressure levels
[in]	nlevq	Number of specific humidity levels
[in]	za	Height of the pressure levels
[in]	zb	Height of the specific humidity levels
[in]	p	Input pressure
[in]	q	Input specific humidity
[in]	vert_interp_ops	Use log(p) for vertical interpolation?
[in]	pseudo_ops	Use pseudo-levels to reduce errors?
[in]	min_temp_grad	Minimum value for the vertical temperature gradient
[in]	nobs	Number of observations
[in]	zobs	Height of the observations
[out]	kmat	K-matrix (Jacobian)

Definition at line 392 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ ufo_gnssro_refmetoffice_simobs_ad()

subroutine ufo_gnssro_refmetoffice_tlad_mod::ufo_gnssro_refmetoffice_simobs_ad	(	class(ufo_gnssro_refmetoffice_tlad), intent(in)	self,
		type(ufo_geovals), intent(inout)	geovals,
		real(kind_real), dimension(:), intent(in)	hofx,
		type(c_ptr), intent(in), value	obss
	)

Given an increment to the observation, find the equivalent increment to the model state.

ufo_gnssro_refmetoffice_simobs_ad

Check that set trajectory has previously been called.
Get the geovals for the model increment, and allocate these if they have not already been allocated.
For each observation calculate the observation increment using the K-matrix

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Parameters

[in]	self	Object which is being used to transfer information
[in,out]	geovals	Calculated perturbations to model state
[in]	hofx	Increment to the observations
[in]	obss	Input - the observations

Definition at line 274 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Here is the call graph for this function:

◆ ufo_gnssro_refmetoffice_simobs_tl()

subroutine ufo_gnssro_refmetoffice_tlad_mod::ufo_gnssro_refmetoffice_simobs_tl	(	class(ufo_gnssro_refmetoffice_tlad), intent(in)	self,
		type(ufo_geovals), intent(in)	geovals,
		real(kind_real), dimension(:), intent(inout)	hofx,
		type(c_ptr), intent(in), value	obss
	)

Given an increment to the model state, calculate an increment to the observation.

ufo_gnssro_refmetoffice_simobs_tl

Check that set trajectory has been previously called.
Get the geovals for the increment.
For each observation apply the K-matrix to calculate the increment to the observation.

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Parameters

[in]	self	Object which is being used to transfer information
[in]	geovals	Model perturbations
[in,out]	hofx	Increment to the observations
[in]	obss	Input - the observations

Definition at line 197 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Here is the call graph for this function:

◆ ufo_gnssro_refmetoffice_tlad_delete()

subroutine ufo_gnssro_refmetoffice_tlad_mod::ufo_gnssro_refmetoffice_tlad_delete ( class(ufo_gnssro_refmetoffice_tlad), intent(inout) self )

Tidy up the variables that are used for passing information.

ufo_gnssro_refmetoffice_tlad_delete

Set lengths to zero, and deallocate K-matrix

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Definition at line 356 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

◆ ufo_gnssro_refmetoffice_tlad_settraj()

subroutine ufo_gnssro_refmetoffice_tlad_mod::ufo_gnssro_refmetoffice_tlad_settraj	(	class(ufo_gnssro_refmetoffice_tlad), intent(inout)	self,
		type(ufo_geovals), intent(in)	geovals,
		type(c_ptr), intent(in), value	obss
	)

Set up the K-matrix for (Jacobian) for the Met Office's GNSS-RO refractivity operator.

ufo_gnssro_refmetoffice_tlad_settraj

It is necessary to run this routine before calling the TL or AD routines.
Get the geovals specifying the state around which to linearise, flipping the vertical order if required.
Call the helper function to calculate the K-matrix for each observation.

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Parameters

[in,out]	self	The object that we use to save data in
[in]	geovals	The input geovals
[in]	obss	The input observations

Definition at line 102 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Here is the call graph for this function:

◆ ufo_gnssro_refmetoffice_tlad_setup()

subroutine ufo_gnssro_refmetoffice_tlad_mod::ufo_gnssro_refmetoffice_tlad_setup	(	class(ufo_gnssro_refmetoffice_tlad), intent(inout)	self,
		logical(c_bool), intent(in)	vert_interp_ops,
		logical(c_bool), intent(in)	pseudo_ops,
		real(c_float), intent(in)	min_temp_grad
	)

Set up the Met Office GNSS-RO refractivity TL/AD.

ufo_gnssro_refmetoffice_tlad_setup

Get the optional settings for the forward model and its linear, and save them in the object so that they can be used in the code.

Author: Neill Bowler (Met Office)

Date: 26 May 2021

Definition at line 72 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Variable Documentation

◆ max_string

integer, parameter ufo_gnssro_refmetoffice_tlad_mod::max_string =800

Definition at line 36 of file ufo_gnssro_refmetoffice_tlad_mod.F90.

Data Types

Functions/Subroutines

Variables

Detailed Description

Function/Subroutine Documentation

◆ jacobian_interface()

◆ ufo_gnssro_refmetoffice_simobs_ad()

◆ ufo_gnssro_refmetoffice_simobs_tl()

◆ ufo_gnssro_refmetoffice_tlad_delete()

◆ ufo_gnssro_refmetoffice_tlad_settraj()

◆ ufo_gnssro_refmetoffice_tlad_setup()

Variable Documentation

◆ max_string