CODEWITHSUNDEEP
segmentation-faultfortranfortran90
Paulo Santiago
Paulo Santiago

Reputation: 43

Segmentation fault on logical variable in Fortran

I'm trying to run the subroutine 'condensation' found here (http://mesonh.aero.obs-mip.fr/chaboureau/PUB/NCL/), thus I wrote a 'main' program in order to initialize the arrays and call the subroutine.

I did 'print' statements in the 'condensation' subroutine to find what was wrong, and I found that the problem (segmentation fault error) occurs on every mention of the logical variable 'LUSERI'.

But, I don't know why.

In the main program, I wrote:

program main
logical :: luseri
...
luseri = .true.
...
call condensation(...,luseri)
end program main

('LUSERI' is the last argument in the subroutine)

Everything seems to be ok: the variable declaration and assignment in the main program, the declaration in the subroutine 'condensation' and its mention.

Here is the 'main' program (main_cst.f90) I wrote:

program main_cst
implicit none

integer, parameter :: klon = 8  ! horizontal dimension
integer, parameter :: klev = 28 ! vertical dimension
integer, parameter :: kidia = 1 ! value of the first point in x; default=1
integer, parameter :: kfdia = 8 ! value of the last point in x; default=KLON
integer, parameter :: kbdia = 1 ! vert. comp. start at KBDIA that is at least 1
integer, parameter :: ktdia = 1 ! vert. comp. can be limited to KLEV + 1 - KTDIA
                                ! default=1                                  
logical            :: luseri    ! logical switch to compute both liquid
                                ! and solid condensate (LUSERI=.TRUE.)
                                ! or only liquid condensate (LUSERI=.FALSE.)
real, dimension(klon,klev) :: ppabs  ! pressure (Pa)
real, dimension(klon,klev) :: pzz    ! height of model levels (m)
real, dimension(klon,klev) :: pt     ! grid scale T  (K)
real, dimension(klon,klev) :: prv    ! grid scale water vapor mixing ratio (kg/kg)
real, dimension(klon,klev) :: pmflx  ! convective mass flux (kg/(s m^2))
real, dimension(klon,klev) :: prc    ! grid scale r_c mixing ratio (kg/kg)
real, dimension(klon,klev) :: pri    ! grid scale r_i (kg/kg)

integer :: kx ! horizontal loop counter

do kx = kidia, kfdia
   ppabs(kx,:)=(/1000, 975, 950, 925, 900, 875, 850, 825, 800, 775, 750, 725, 700, &
            675, 650, 600, 500, 300, 250, 200, 150, 100, 70, 50, 30, 20, 10, 3/)
   pzz(kx,:)=(/111.31, 333.80, 561.25, 794.10, 1032.35, 1276.02, 1525.42, &
           1780.62, 2041.78, 2309.21, 2583.33, 2864.61, 3153.70, 3451.53, &
           3758.06, 4397.20, 5770.03, 9485.69, 10710.62, 12141.42, 13897.52, &
          16283.89, 18361.25, 20367.34, 23523.77, 26158.47, 30748.48, 38877.28/)
   pt(kx,:)=(/300.81, 299.97, 299.63, 298.31, 296.93, 295.81, 294.07, 292.04, &
          290.02, 287.85, 285.50, 283.34, 281.39, 279.55, 278.16, 275.40, &
          267.42, 239.07, 229.57, 219.98, 208.53, 197.10, 193.60, 200.84, &
          213.77, 221.35, 230.21, 231.10/)
   prv(kx,:)=(/0.012570000, 0.012460000, 0.011830000, 0.011390000, 0.010560000, &
           0.009922000, 0.009529000, 0.009226000, 0.009008000, 0.008739000, &
           0.008411000, 0.007836000, 0.007077000, 0.006153000, 0.004703000, &
           0.002451000, 0.000774100, 0.000107400, 0.000056510, 0.000031610, &
           0.000008088, 0.000004136, 0.000002686, 0.000002901, 0.000003876, &
           0.000004562, 0.000004388, 0.000007886/)
end do

pmflx(:,:) = 0.0
prc(:,:) = 0.0
pri(:,:) = 0.0
luseri = .true.

call condensation(klon, klev, kidia, kfdia, kbdia, ktdia, ppabs, pzz, pt, prv, &
                  pmflx, prc, pri, luseri)

end program main_cst

The condensation subroutine (condensation.f90) can be found at http://mesonh.aero.obs-mip.fr/chaboureau/PUB/NCL/:

!     ######spl
    SUBROUTINE CONDENSATION( KLON, KLEV, KIDIA, KFDIA, KBDIA, KTDIA,            &
                           PPABS, PZZ, PT, PRV, PRC, PRI, PMFLX, PCLDFR, LUSERI )
!   ############################################################################
!
!!
!!    PURPOSE
!!    -------
!!**  Routine to diagnose cloud fraction and liquid and ice condensate mixing ratios
!!     
!!    
!!**  METHOD
!!    ------
!!    Based on the large-scale fields of temperature, water vapor, and possibly
!!    liquid and solid condensate, the conserved quantities r_t and h_l are constructed 
!!    and then fractional cloudiness, liquid and solid condensate is diagnosed.
!!
!!    The total variance is parameterized as the sum of  stratiform/turbulent variance 
!!    and a convective variance.
!!    The turbulent variance is parameterized as a function of first-order moments, and
!!    the convective variance is modelled as a function of the convective mass flux (units kg/s m^2)
!!    as provided by a mass flux convection scheme.
!!
!!    Nota: if the host model does not use prognostic values for liquid and solid condensate
!!    or does not provide a convective mass flux, put all these values to zero.
!!    Also, it is supposed that vertical model levels are numbered from
!!    1 to KLEV, where 1 is the first model level above the surface
!!      
!!     
!!
!!    EXTERNAL
!!    --------
!!      INI_CST
!!
!!    IMPLICIT ARGUMENTS
!!    ------------------
!!      Module MODD_CST       : contains physical constants
!!
!!    REFERENCE
!!    ---------
!!      Chaboureau J.P. and P. Bechtold (J. Atmos. Sci. 2002)
!!
!!    AUTHOR
!!    ------
!!      P. BECHTOLD       * Laboratoire d'Aerologie *
!!
!!    MODIFICATIONS
!!    -------------
!!      Original    13/06/2001
!!      modified    20/03/2002 : add convective Sigma_s and improve turbulent
!!                               length-scale in boundary-layer and near tropopause
!!
!-------------------------------------------------------------------------------
!
!*       0.    DECLARATIONS
!              ------------
!
USE MODD_CST
!
IMPLICIT NONE
!
!*       0.1   Declarations of dummy arguments :
!
!
INTEGER,                    INTENT(IN)   :: KLON    ! horizontal dimension
INTEGER,                    INTENT(IN)   :: KLEV    ! vertical dimension
INTEGER,                    INTENT(IN)   :: KIDIA   ! value of the first point in x
                                                    ! default=1
INTEGER,                    INTENT(IN)   :: KFDIA   ! value of the last point in x
                                                    ! default=KLON
INTEGER,                    INTENT(IN)   :: KBDIA   ! vertical  computations start at
!                                                   ! KBDIA that is at least 1
INTEGER,                    INTENT(IN)   :: KTDIA   ! vertical computations can be
                                                    ! limited to KLEV + 1 - KTDIA
                                                    ! default=1
REAL, DIMENSION(KLON,KLEV), INTENT(IN)    :: PPABS  ! pressure (Pa)
REAL, DIMENSION(KLON,KLEV), INTENT(IN)    :: PZZ    ! height of model levels (m)
REAL, DIMENSION(KLON,KLEV), INTENT(IN)    :: PT     ! grid scale T  (K)
REAL, DIMENSION(KLON,KLEV), INTENT(IN)    :: PRV    ! grid scale water vapor mixing ratio (kg/kg)
LOGICAL                                   :: LUSERI ! logical switch to compute both
                            ! liquid and solid condensate (LUSERI=.TRUE.)
                            ! or only liquid condensate (LUSERI=.FALSE.)
REAL, DIMENSION(KLON,KLEV), INTENT(IN)    :: PMFLX  ! convective mass flux (kg/(s m^2))
REAL, DIMENSION(KLON,KLEV), INTENT(INOUT) :: PRC    ! grid scale r_c mixing ratio (kg/kg)
REAL, DIMENSION(KLON,KLEV), INTENT(INOUT) :: PRI    ! grid scale r_i (kg/kg)
REAL, DIMENSION(KLON,KLEV), INTENT(OUT)   :: PCLDFR ! fractional cloudiness (between 0 and 1)
!
!                        
!*       0.2   Declarations of local variables :
!
INTEGER  :: JI, JK, JKT, JKP, JKM     ! loop index
REAL, DIMENSION(KLON,KLEV) :: ZTLK, ZRT       ! work arrays for T_l, r_t 
REAL, DIMENSION(KLON,KLEV) :: ZL              ! length-scale
INTEGER, DIMENSION(KLON)   :: ITPL    ! top levels of tropopause/highest inversion
REAL, DIMENSION(KLON)      :: ZTMIN   ! min Temp. related to ITPL
!
REAL :: ZTEMP, ZLV, ZLS, ZTL, ZPV, ZQSL, ZPIV, ZQSI, ZFRAC, ZCOND, ZCPD ! thermodynamics
REAL :: ZLL, DZZ, ZZZ ! length scales
REAL :: ZAH, ZA, ZB, ZSBAR, ZQ1, ZSIGMA, ZDRW, ZDTL ! related to computation of Sig_s
REAL :: ZSIG_CONV                                   ! convective part of Sig_s
!
!*       0.3  Definition of constants :  
!
!-------------------------------------------------------------------------------
!
REAL :: ZL0     = 600.        ! tropospheric length scale
                              ! changed to 600 m instead of 900 m to give a consistent
                              ! value (linear increase) in general 500 m deep oceanic
                              ! mixed layer - but could be put back to 900 m if wished
REAL :: ZCSIGMA = 0.2         ! constant in sigma_s parameterization
REAL :: ZCSIG_CONV = 0.30E-2  ! scaling factor for ZSIG_CONV as function of mass flux 
!
!-------------------------------------------------------------------------------
!

CALL INI_CST     ! Initialize thermodynamic constants in module MODD_CST

PCLDFR(:,:) = 0. ! Initialize values

JKT = KLEV+1-KTDIA
DO JK=KBDIA,JKT
DO JI=KIDIA,KFDIA
   ZTEMP  = PT(JI,JK)
    !latent heat of vaporisation/sublimation
   ZLV    = XLVTT + ( XCPV - XCL ) * ( ZTEMP - XTT )
   ZLS    = XLSTT + ( XCPV - XCI ) * ( ZTEMP - XTT )

    !store temperature at saturation and total water mixing ratio
   ZRT(JI,JK)   = PRV(JI,JK) + PRC(JI,JK) + PRI(JI,JK)
   ZCPD         = XCPD + ZRT(JI,JK) * XCPV
   ZTLK(JI,JK)  = ZTEMP - ZLV*PRC(JI,JK)/ZCPD - ZLS*PRI(JI,JK)/ZCPD
END DO
END DO

!-------------------------------------------------------------------------------
! Determine tropopause/inversion  height from minimum temperature

ITPL(:)  = KBDIA+1
ZTMIN(:) = 400.
DO JK = KBDIA+1,JKT-1
   DO JI=KIDIA,KFDIA 
         IF ( PT(JI,JK) < ZTMIN(JI) ) THEN
              ZTMIN(JI) = PT(JI,JK)
              ITPL(JI) = JK
         END IF
   END DO
END DO

! Set the mixing length scale - used for computing the "turbulent part" of Sigma_s

ZL(:,KBDIA) = 20.
DO JK = KBDIA+1,JKT
DO JI=KIDIA,KFDIA 
      ! free troposphere
   ZL(JI,JK) = ZL0
   JKP = ITPL(JI)
   ZZZ =  PZZ(JI,JK) -  PZZ(JI,KBDIA)
      ! approximate length for boundary-layer : linear increase
   IF ( ZL0 > ZZZ )  ZL(JI,JK) = ZZZ
      ! gradual decrease of length-scale near and above tropopause/top inversion
   IF ( ZZZ > 0.9*(PZZ(JI,JKP)-PZZ(JI,KBDIA)) ) &
        ZL(JI,JK) = .6 * ZL(JI,JK-1) 
END DO
END DO
!-------------------------------------------------------------------------------


DO JK=KBDIA+1,JKT-1
   JKP=JK+1
   JKM=JK-1
DO JI=KIDIA,KFDIA
   ZTEMP  = PT(JI,JK)
    !latent heat of vaporisation/sublimation
   ZLV    = XLVTT + ( XCPV - XCL ) * ( ZTEMP - XTT )
   ZLS    = XLSTT + ( XCPV - XCI ) * ( ZTEMP - XTT )

   ZCPD   = XCPD + ZRT(JI,JK) * XCPV
    !temperature at saturation
   ZTL    = ZTEMP - ZLV*PRC(JI,JK)/ZCPD - ZLS*PRI(JI,JK)/ZCPD
    !saturated water vapor mixing ratio over liquid water
   ZPV    = EXP( XALPW - XBETAW / ZTL - XGAMW * LOG( ZTL ) )
   ZQSL   = XRD / XRV * ZPV / ( PPABS(JI,JK) - ZPV )

    !saturated water vapor mixing ratio over ice
   ZPIV   = EXP( XALPI - XBETAI / ZTL - XGAMI * LOG( ZTL ) )
   ZQSI   = XRD / XRV * ZPIV / ( PPABS(JI,JK) - ZPIV )

    !interpolate between liquid and solid as function of temperature
    ! glaciation interval is specified here to 20 K
   ZFRAC = ( ZTL  - 250.16 ) / ( XTT - 250.16 )  ! liquid/solid fraction
   ZFRAC = MAX( 0., MIN(1., ZFRAC ) )
   ZFRAC = ZFRAC * ZFRAC
   IF(.NOT. LUSERI) ZFRAC=1.
   ZQSL = ( 1. - ZFRAC ) * ZQSI + ZFRAC * ZQSL
   ZLV  = ( 1. - ZFRAC ) * ZLS  + ZFRAC * ZLV

    !coefficients a and b
   ZAH  = ZLV * ZQSL / ( XRV * ZTL**2 )
   ZA   = 1. / ( 1. + ZLV/ZCPD * ZAH )
   ZB   = ZAH * ZA

    !parameterize Sigma_s with first_order closure

   DZZ    =  PZZ(JI,JKP)  - PZZ(JI,JKM)
   ZDRW   =  ZRT(JI,JKP)  - ZRT(JI,JKM)
   ZDTL   =  ZTLK(JI,JKP) - ZTLK(JI,JKM) + XG/ZCPD * DZZ
   ZLL    =  ZL(JI,JK)

   ZSIG_CONV = ZCSIG_CONV * PMFLX(JI,JK) / ZA ! standard deviation due to convection
   ZSIGMA =  SQRT( MAX( 1.E-25, ZCSIGMA*ZCSIGMA* ZLL*ZLL/(DZZ*DZZ) * ( &
             ZA*ZA*ZDRW*ZDRW - 2.*ZA*ZB*ZDRW*ZDTL + ZB*ZB*ZDTL*ZDTL  ) &
                                       + ZSIG_CONV * ZSIG_CONV ) )
    !zsigma should be of order 4.e-4 in lowest 5 km of atmosphere
   ZSIGMA = MAX( ZSIGMA, 1.E-12 )

    !normalized saturation deficit
   ZSBAR = ZA * ( ZRT(JI,JK) - ZQSL )
   ZQ1   = ZSBAR / ZSIGMA 

    !cloud fraction
   PCLDFR(JI,JK) = MAX( 0., MIN(1.,0.5+0.36*ATAN(1.55*ZQ1)) )

    !total condensate
   IF (ZQ1 > 0. .AND. ZQ1 <= 2. ) THEN
      ZCOND = EXP(-1.)+.66*ZQ1+.086*ZQ1*ZQ1
   ELSE IF (ZQ1 > 2.) THEN
      ZCOND = ZQ1
   ELSE
      ZCOND = EXP( 1.2*ZQ1-1 )
   END IF
   ZCOND = ZCOND * ZSIGMA

   if ( zcond<1.e-6) then
       zcond = 0.
       pcldfr(ji,jk) = 0.
   end if

   PRC(JI,JK) = ZFRAC * ZCOND ! liquid condensate
   IF (LUSERI) THEN
      PRI(JI,JK) = (1.-ZFRAC) * ZCOND   ! solid condensate
   END IF


    ! compute s'rl'/Sigs^2
    ! used in w'rl'= w's' * s'rl'/Sigs^2
!  PSIGRC(JI,JK) = PCLDFR(JI,JK)   ! Gaussian relation 

END DO
END DO
!
END SUBROUTINE CONDENSATION

There are two more files, also found at http://mesonh.aero.obs-mip.fr/chaboureau/PUB/NCL/.

ini_cst.f90:

!     ######spl
      SUBROUTINE INI_CST
!     ##################
!
!!****  *INI_CST * - routine to initialize the constants modules 
!!
!!    PURPOSE
!!    -------
!       The purpose of this routine is to initialize  the constants
!     stored in  modules MODD_CST
!      
!
!!**  METHOD
!!    ------
!!      The thermodynamic constants are set to their numerical values 
!!     
!!
!!    EXTERNAL
!!    --------
!!
!!    IMPLICIT ARGUMENTS
!!    ------------------
!!      Module MODD_CST       : contains physical constants
!!
!!    REFERENCE
!!    ---------
!!      Chaboureau J.P. and P. Bechtold (J. Atmos. Sci. 2002)
!!      
!!
!!    AUTHOR
!!    ------
!!      P. BECHTOLD       * Laboratoire d'Aerologie *
!!
!!    MODIFICATIONS
!!    -------------
!!      Original    13/06/2001
!!      modified    20/03/2002 : add convective Sigma_s and improve turbulent
!!                               length-scale in boundary-layer and near tropopause
!-------------------------------------------------------------------------------
!
!*       0.    DECLARATIONS
!              ------------
!
USE MODD_CST
!
IMPLICIT NONE
!  
!-------------------------------------------------------------------------------
!
!*       1.    Set the fundamental thermodynamical constants
!              these have the same values (not names) as in ARPEGE IFS 
!              -------------------------------------------------------
!
!
XP00   = 1.E5        ! reference pressure
XPI    = 3.141592654 ! Pi
 XG    = 9.80665     ! gravity constant
XMD    = 28.9644E-3  ! molecular weight of dry air
XMV    = 18.0153E-3  ! molecular weight of water vapor
XRD    = 287.05967   ! gaz constant for dry air
XRV    = 461.524993  ! gaz constant for water vapor
XCPD   = 1004.708845 ! specific heat of dry air
XCPV   = 1846.1      ! specific heat of water vapor
XRHOLW = 1000.       ! density of liquid water
XCL    = 4218.       ! specific heat of liquid water
XCI    = 2106.       ! specific heat of ice
XTT    = 273.16      ! triple point temperature
XLVTT  = 2.5008E6    ! latent heat of vaporisation at XTT
XLSTT  = 2.8345E6    ! latent heat of sublimation at XTT 
XLMTT  = 0.3337E6    ! latent heat of melting at XTT
XESTT  = 611.14      ! saturation pressure at XTT
XALPW  = 60.22416    ! constants in saturation pressure over liquid water
XBETAW = 6822.459384
XGAMW  = 5.13948
XALPI  = 32.62116    ! constants in saturation pressure over ice
XBETAI = 6295.421
XGAMI  = 0.56313
!
!
END SUBROUTINE INI_CST

!     ######spl
      MODULE MODD_CST
!     ###############
!
IMPLICIT NONE
!
REAL, SAVE :: XP00   ! reference pressure
REAL, SAVE :: XPI    ! Pi
REAL, SAVE ::  XG    ! gravity constant
REAL, SAVE :: XMD    ! molecular weight of dry air
REAL, SAVE :: XMV    ! molecular weight of water vapor
REAL, SAVE :: XRD    ! gaz constant for dry air
REAL, SAVE :: XRV    ! gaz constant for water vapor
REAL, SAVE :: XCPD   ! specific heat of dry air
REAL, SAVE :: XCPV   ! specific heat of water vapor
REAL, SAVE :: XRHOLW ! density of liquid water
REAL, SAVE :: XCL    ! specific heat of liquid water
REAL, SAVE :: XCI    ! specific heat of ice
REAL, SAVE :: XTT    ! triple point temperature
REAL, SAVE :: XLVTT  ! latent heat of vaporisation at XTT
REAL, SAVE :: XLSTT  ! latent heat of sublimation at XTT 
REAL, SAVE :: XLMTT  ! latent heat of melting at XTT
REAL, SAVE :: XESTT  ! saturation pressure at XTT
REAL, SAVE :: XALPW  ! constants in saturation pressure over liquid water
REAL, SAVE :: XBETAW 
REAL, SAVE :: XGAMW 
REAL, SAVE :: XALPI  ! constants in saturation pressure over ice
REAL, SAVE :: XBETAI 
REAL, SAVE :: XGAMI 
!
END MODULE MODD_CST

modd_cst.f90:

!     ######spl
      MODULE MODD_CST
!     ###############
!
IMPLICIT NONE
!
REAL, SAVE :: XP00   ! reference pressure
REAL, SAVE :: XPI    ! Pi
REAL, SAVE ::  XG    ! gravity constant
REAL, SAVE :: XMD    ! molecular weight of dry air
REAL, SAVE :: XMV    ! molecular weight of water vapor
REAL, SAVE :: XRD    ! gaz constant for dry air
REAL, SAVE :: XRV    ! gaz constant for water vapor
REAL, SAVE :: XCPD   ! specific heat of dry air
REAL, SAVE :: XCPV   ! specific heat of water vapor
REAL, SAVE :: XRHOLW ! density of liquid water
REAL, SAVE :: XCL    ! specific heat of liquid water
REAL, SAVE :: XCI    ! specific heat of ice
REAL, SAVE :: XTT    ! triple point temperature
REAL, SAVE :: XLVTT  ! latent heat of vaporisation at XTT
REAL, SAVE :: XLSTT  ! latent heat of sublimation at XTT 
REAL, SAVE :: XLMTT  ! latent heat of melting at XTT
REAL, SAVE :: XESTT  ! saturation pressure at XTT
REAL, SAVE :: XALPW  ! constants in saturation pressure over liquid water
REAL, SAVE :: XBETAW 
REAL, SAVE :: XGAMW 
REAL, SAVE :: XALPI  ! constants in saturation pressure over ice
REAL, SAVE :: XBETAI 
REAL, SAVE :: XGAMI 
!
END MODULE MODD_CST

This is all the code I'm working with.

I use the command to compile:

g95 modd_cst.f90 ini_cst.f90 condensation.f90 main_cst.f90 -o cst.g95

Upvotes: 0

Views: 370

Answers (1)

horseshoe
horseshoe

Reputation: 1477

You seem to have a mismatch between what you put in the call and what the subroutine gets

call condensation(klon, klev, kidia, kfdia, kbdia, ktdia, ppabs, pzz, pt, prv, 
                  pmflx, prc, pri, luseri)

and the subroutine wants

 SUBROUTINE CONDENSATION( KLON, KLEV, KIDIA, KFDIA, KBDIA, KTDIA,            &
                           PPABS, PZZ, PT, PRV, PRC, PRI, PMFLX, PCLDFR, LUSERI )

so in the call its prv pmflx prc pri in the subroutine its prv prc pri pmflx. Furthermore you are missing one variable that you pass to the subroutine. Fortran is sensitive to the order of the variables that you pass on. Without checking the rest of the code I assume that should solve the problem.

Upvotes: 1

Related Questions