diverr_mod Module Reference

Calculates the local divergence error. More...

Functions/Subroutines
type(component_descriptor_type) function, public	diverr_get_descriptor ()
	Descriptor of this component for registration. More...
subroutine	init_callback (current_state)
	The initialisation callback will allocate memory for the P field and initialise it. More...
subroutine	timestep_callback (current_state)
	Called each timestep this will initialise P and update the local divergence error for the current column. More...
subroutine	finalisation_callback (current_state)
	Called at finalisation this will deallocate the P field as the model shuts down. More...
subroutine	find_max_divergence_error (current_state)
	Finds the maximum divergence error locally and writes this into the local variable. More...
subroutine	handle_forward_stepping (current_state)
	Handles the calculation of P when the model is forward stepping. More...
subroutine	handle_centred_stepping (current_state)
	Handles the calculation of P when the model is centred stepping. More...
subroutine	calculate_p (current_state, p, u, v, w, timec, y_local, x_local)
	Calculates P based upon flow fields and the stepping. More...

Detailed Description

Calculates the local divergence error.

Function/Subroutine Documentation

calculate_p()

subroutine diverr_mod::calculate_p ( type(model_state_type), intent(inout), target  current_state, type(prognostic_field_type), intent(inout)  p, type(prognostic_field_type), intent(inout)  u, type(prognostic_field_type), intent(inout)  v, type(prognostic_field_type), intent(inout)  w, real(kind=default_precision), intent(in)  timec, integer, intent(in)  y_local, integer, intent(in)  x_local  )

private

Calculates P based upon flow fields and the stepping.

Parameters

current_state	The current model state
p	The P field to update
u	The U flow field
v	The V flow field
w	The W flow field
timec	Based upon model dtm this depends on the stepping selected
y_local	Local Y dimension index
x_local	Local X dimension index

Definition at line 120 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
    type(prognostic_field_type), intent(inout) :: u, v, w, p
    real(kind=default_precision), intent(in) :: timec
    integer, intent(in) :: y_local, x_local
    
    integer :: k
    
    p%data(:, y_local, x_local)=0.0_default_precision
    do k=2,current_state%local_grid%size(z_index)
#ifdef U_ACTIVE
      p%data(k, y_local, x_local)= p%data(k, y_local, x_local)+ &
           current_state%global_grid%configuration%horizontal%cx*(u%data(k, y_local, x_local)-u%data(k, y_local, x_local-1))
#endif
#ifdef V_ACTIVE
      p%data(k, y_local, x_local)= p%data(k, y_local, x_local)+ &
           current_state%global_grid%configuration%horizontal%cy*(v%data(k, y_local, x_local)-v%data(k, y_local-1, x_local))
#endif
#ifdef W_ACTIVE
      p%data(k, y_local, x_local)= p%data(k, y_local, x_local)+ &
           4.0_default_precision*(current_state%global_grid%configuration%vertical%tzc2(k)*w%data(k, y_local, x_local)-&
           current_state%global_grid%configuration%vertical%tzc1(k)* w%data(k-1, y_local, x_local))
#endif
      p%data(k, y_local, x_local)=p%data(k, y_local, x_local) * timec
    end do

Here is the caller graph for this function:

diverr_get_descriptor()

type(component_descriptor_type) function, public diverr_mod::diverr_get_descriptor

Descriptor of this component for registration.

Returns

The diverr component descriptor

Definition at line 19 of file diverr.F90.

    diverr_get_descriptor%name="diverr"
    diverr_get_descriptor%version=0.1
    diverr_get_descriptor%initialisation=>init_callback
    diverr_get_descriptor%timestep=>timestep_callback
    diverr_get_descriptor%finalisation=>finalisation_callback

Here is the call graph for this function:

finalisation_callback()

subroutine diverr_mod::finalisation_callback ( type(model_state_type), intent(inout), target  current_state )

private

Called at finalisation this will deallocate the P field as the model shuts down.

Parameters

current_state

The current model state

Definition at line 61 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    current_state%p%active=.false.
    deallocate(current_state%p%data)

Here is the caller graph for this function:

find_max_divergence_error()

subroutine diverr_mod::find_max_divergence_error ( type(model_state_type), intent(inout), target  current_state )

private

Finds the maximum divergence error locally and writes this into the local variable.

Parameters

current_state

The current model state

Definition at line 70 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    integer :: k
 
    if (current_state%first_timestep_column) current_state%local_divmax=0.0_default_precision
 
    do k=2,current_state%local_grid%size(z_index)
      if (abs(current_state%p%data(k, current_state%column_local_y, current_state%column_local_x)) .gt. &
           current_state%local_divmax) then
        current_state%local_divmax=abs(current_state%p%data(k,current_state%column_local_y, current_state%column_local_x))
      end if
    end do

Here is the caller graph for this function:

handle_centred_stepping()

subroutine diverr_mod::handle_centred_stepping ( type(model_state_type), intent(inout), target  current_state )

private

Handles the calculation of P when the model is centred stepping.

Parameters

current_state

The current model state

Definition at line 100 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    real(kind=default_precision) :: timec
 
    timec=1.0_default_precision/(2.0_default_precision*current_state%dtm)
 
    call calculate_p(current_state, current_state%p, current_state%zu, current_state%zv, current_state%zw, timec, &
         current_state%column_local_y, current_state%column_local_x)

Here is the call graph for this function:

Here is the caller graph for this function:

handle_forward_stepping()

subroutine diverr_mod::handle_forward_stepping ( type(model_state_type), intent(inout), target  current_state )

private

Handles the calculation of P when the model is forward stepping.

Parameters

current_state

The current model state

Definition at line 87 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    real(kind=default_precision) :: timec
 
    timec=1.0_default_precision/current_state%dtm
   
    call calculate_p(current_state, current_state%p, current_state%u, current_state%v, current_state%w, timec, &
         current_state%column_local_y, current_state%column_local_x)

Here is the call graph for this function:

Here is the caller graph for this function:

init_callback()

subroutine diverr_mod::init_callback ( type(model_state_type), intent(inout), target  current_state )

private

The initialisation callback will allocate memory for the P field and initialise it.

Parameters

current_state

The current model state

Definition at line 29 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    if (.not. allocated(current_state%p%data)) then
      ! If we are loading from a checkpoint file then this is already present
      allocate(current_state%p%data(current_state%local_grid%size(z_index) + current_state%local_grid%halo_size(z_index) * 2, &
           current_state%local_grid%size(y_index) + current_state%local_grid%halo_size(y_index) * 2, &
           current_state%local_grid%size(x_index) + current_state%local_grid%halo_size(x_index) * 2))
      current_state%p%data=0.0_default_precision
      current_state%p%active=.true.
    end if

Here is the caller graph for this function:

timestep_callback()

subroutine diverr_mod::timestep_callback ( type(model_state_type), intent(inout), target  current_state )

private

Called each timestep this will initialise P and update the local divergence error for the current column.

Parameters

current_state

The current model state

Definition at line 44 of file diverr.F90.

    type(model_state_type), target, intent(inout) :: current_state
 
    current_state%p%data(:, current_state%column_local_y, current_state%column_local_x) = 0.0_default_precision
 
    if (current_state%halo_column) return
    
    if (current_state%field_stepping == forward_stepping) then
      call handle_forward_stepping(current_state)
    else
      call handle_centred_stepping(current_state)
    end if    
    call find_max_divergence_error(current_state)

Here is the call graph for this function:

Here is the caller graph for this function: