module intrwmod
!$$$ module documentation block
! . . . .
! module: intrwmod module for intrw and its tangent linear intrw_tl
! prgmmr:
!
! abstract: module for intrw and its tangent linear intrw_tl
!
! program history log:
! 2005-05-13 Yanqiu zhu - wrap intrw and its tangent linear intrw_tl into one module
! 2005-11-16 Derber - remove interfaces
! 2008-11-26 Todling - remove intrw_tl; add interface back
! 2009-08-13 lueken - update documentation
!
! subroutines included:
! sub intrw_
!
! variable definitions:
!
! attributes:
! language: f90
! machine:
!
!$$$ end documentation block
implicit none
PRIVATE
PUBLIC intrw
interface intrw; module procedure &
intrw_
end interface
contains
subroutine intrw_(rwhead,ru,rv,su,sv)
!$$$ subprogram documentation block
! . . . .
! subprogram: intrw apply nonlin qc operator for radar winds
! prgmmr: derber org: np23 date: 1991-02-26
!
! abstract: apply observation operator for radar winds
! with nonlinear qc operator
!
! program history log:
! 1991-02-26 derber
! 1999-11-22 yang
! 2004-08-02 treadon - add only to module use, add intent in/out
! 2004-10-08 parrish - add nonlinear qc option
! 2005-03-01 parrish - nonlinear qc change to account for inflated obs error
! 2005-04-11 treadon - merge intrw and intrw_qc into single routine
! 2005-08-02 derber - modify for variational qc parameters for each ob
! 2005-09-28 derber - consolidate location and weight arrays
! 2006-07-28 derber - modify to use new inner loop obs data structure
! - unify NL qc
! 2007-02-15 rancic - add foto
! 2007-03-19 tremolet - binning of observations
! 2007-06-05 tremolet - use observation diagnostics structure
! 2007-07-09 tremolet - observation sensitivity
! 2008-01-04 tremolet - Don't apply H^T if l_do_adjoint is false
! 2008-11-28 todlng - turn FOTO optional; changed ptr%time handle
!
! usage: call intw(ru,rv,su,sv)
! input argument list:
! rwhead - obs type pointer to obs structure
! su - current u solution increment
! sv - current v solution increment
! ru
! rv
!
! output argument list:
! ru - u results from observation operator
! rv - v results from observation operator
!
! attributes:
! language: f90
! machine: ibm RS/6000 SP
!
!$$$
use kinds, only: r_kind,i_kind
use constants, only: half,one,tiny_r_kind,cg_term,r3600
use obsmod, only: rw_ob_type,lsaveobsens,l_do_adjoint
use qcmod, only: nlnqc_iter,varqc_iter
use gridmod, only: latlon1n
use jfunc, only: jiter,l_foto,xhat_dt,dhat_dt
implicit none
! Declare passed variables
type(rw_ob_type),pointer ,intent(in ) :: rwhead
real(r_kind),dimension(latlon1n),intent(in ) :: su,sv
real(r_kind),dimension(latlon1n),intent(inout) :: ru,rv
! Declare local varibles
integer(i_kind) j1,j2,j3,j4,j5,j6,j7,j8
! real(r_kind) penalty
real(r_kind) val,valu,valv,w1,w2,w3,w4,w5,w6,w7,w8
real(r_kind) cg_rw,p0,grad,wnotgross,wgross,time_rw,pg_rw
type(rw_ob_type), pointer :: rwptr
rwptr => rwhead
do while (associated(rwptr))
j1=rwptr%ij(1)
j2=rwptr%ij(2)
j3=rwptr%ij(3)
j4=rwptr%ij(4)
j5=rwptr%ij(5)
j6=rwptr%ij(6)
j7=rwptr%ij(7)
j8=rwptr%ij(8)
w1=rwptr%wij(1)
w2=rwptr%wij(2)
w3=rwptr%wij(3)
w4=rwptr%wij(4)
w5=rwptr%wij(5)
w6=rwptr%wij(6)
w7=rwptr%wij(7)
w8=rwptr%wij(8)
! Forward mode l
val=(w1* su(j1)+w2* su(j2)+w3* su(j3)+w4* su(j4)+ &
w5* su(j5)+w6* su(j6)+w7* su(j7)+w8* su(j8))*rwptr%cosazm+ &
(w1* sv(j1)+w2* sv(j2)+w3* sv(j3)+w4* sv(j4)+ &
w5* sv(j5)+w6* sv(j6)+w7* sv(j7)+w8* sv(j8))*rwptr%sinazm
if ( l_foto ) then
time_rw=rwptr%time*r3600
val=val+ &
((w1*xhat_dt%u(j1)+w2*xhat_dt%u(j2)+ &
w3*xhat_dt%u(j3)+w4*xhat_dt%u(j4)+ &
w5*xhat_dt%u(j5)+w6*xhat_dt%u(j6)+ &
w7*xhat_dt%u(j7)+w8*xhat_dt%u(j8))*rwptr%cosazm+ &
(w1*xhat_dt%v(j1)+w2*xhat_dt%v(j2)+ &
w3*xhat_dt%v(j3)+w4*xhat_dt%v(j4)+ &
w5*xhat_dt%v(j5)+w6*xhat_dt%v(j6)+ &
w7*xhat_dt%v(j7)+w8*xhat_dt%v(j8))*rwptr%sinazm)*time_rw
endif
if (lsaveobsens) then
rwptr%diags%obssen(jiter) = val*rwptr%raterr2*rwptr%err2
else
if (rwptr%luse) rwptr%diags%tldepart(jiter)=val
endif
if (l_do_adjoint) then
if (lsaveobsens) then
grad = rwptr%diags%obssen(jiter)
else
val=val-rwptr%res
! gradient of nonlinear operator
if (nlnqc_iter .and. rwptr%pg > tiny_r_kind .and. &
rwptr%b > tiny_r_kind) then
pg_rw=rwptr%pg*varqc_iter
cg_rw=cg_term/rwptr%b
wnotgross= one-pg_rw
wgross = pg_rw*cg_rw/wnotgross
p0 = wgross/(wgross+exp(-half*rwptr%err2*val**2))
val = val*(one-p0)
endif
grad = val*rwptr%raterr2*rwptr%err2
endif
! Adjoint
valu=rwptr%cosazm*grad
valv=rwptr%sinazm*grad
ru(j1)=ru(j1)+w1*valu
ru(j2)=ru(j2)+w2*valu
ru(j3)=ru(j3)+w3*valu
ru(j4)=ru(j4)+w4*valu
ru(j5)=ru(j5)+w5*valu
ru(j6)=ru(j6)+w6*valu
ru(j7)=ru(j7)+w7*valu
ru(j8)=ru(j8)+w8*valu
rv(j1)=rv(j1)+w1*valv
rv(j2)=rv(j2)+w2*valv
rv(j3)=rv(j3)+w3*valv
rv(j4)=rv(j4)+w4*valv
rv(j5)=rv(j5)+w5*valv
rv(j6)=rv(j6)+w6*valv
rv(j7)=rv(j7)+w7*valv
rv(j8)=rv(j8)+w8*valv
if ( l_foto ) then
valu=valu*time_rw
valv=valv*time_rw
dhat_dt%u(j1)=dhat_dt%u(j1)+w1*valu
dhat_dt%u(j2)=dhat_dt%u(j2)+w2*valu
dhat_dt%u(j3)=dhat_dt%u(j3)+w3*valu
dhat_dt%u(j4)=dhat_dt%u(j4)+w4*valu
dhat_dt%u(j5)=dhat_dt%u(j5)+w5*valu
dhat_dt%u(j6)=dhat_dt%u(j6)+w6*valu
dhat_dt%u(j7)=dhat_dt%u(j7)+w7*valu
dhat_dt%u(j8)=dhat_dt%u(j8)+w8*valu
dhat_dt%v(j1)=dhat_dt%v(j1)+w1*valv
dhat_dt%v(j2)=dhat_dt%v(j2)+w2*valv
dhat_dt%v(j3)=dhat_dt%v(j3)+w3*valv
dhat_dt%v(j4)=dhat_dt%v(j4)+w4*valv
dhat_dt%v(j5)=dhat_dt%v(j5)+w5*valv
dhat_dt%v(j6)=dhat_dt%v(j6)+w6*valv
dhat_dt%v(j7)=dhat_dt%v(j7)+w7*valv
dhat_dt%v(j8)=dhat_dt%v(j8)+w8*valv
endif
endif
rwptr => rwptr%llpoint
end do
return
end subroutine intrw_
end module intrwmod