I have a few "Argument is one type at (1) but is some other type at (2) errors coming up in my program, but I'm not sure how to modify it. Is there some other way of writing the arguments than what I've done?
Ps. I compile in g77. But I also tried in gfortran, where the error that comes up is of the kind "Rank mismatch in argument 'z' at (1) (scalar and rank-1)".
Here are the errors coming up:
geomalb.f: In subroutine `atmsetup':
geomalb.f:1083:
SUBROUTINE ATMSETUP(NLEVEL,Z,RHCH4,FH2,FARGON,TEMP,PRESS,DEN,XMU,
1
geomalb.f:1122: (continued):
SUMT=SUMT+(EFFG(Z(J))*ADEN)*( Z(J-1)-Z(J))*XMU(J)
2
Invalid declaration of or reference to symbol `z' at (2) [initially seen at (1)]
geomalb.f:80: warning:
COLDEN(J)=RHOP*(PRESS(J+1)-PRESS(J))/EFFG(Z(J))
1
geomalb.f:1122: (continued):
SUMT=SUMT+(EFFG(Z(J))*ADEN)*( Z(J-1)-Z(J))*XMU(J)
2
Argument #1 of `effg' is one type at (2) but is some other type at (1) [info -f g77 M GLOBALS]
geomalb.f:80: warning:
COLDEN(J)=RHOP*(PRESS(J+1)-PRESS(J))/EFFG(Z(J))
1
geomalb.f:1124: (continued):
SUMB=SUMB+(EFFG(Z(J))*ADEN)*( Z(J-1)-Z(J))*28.01340
2
Argument #1 of `effg' is one type at (2) but is some other type at (1) [info -f g77 M GLOBALS]
geomalb.f:72: warning:
CALL ATMSETUP(NLEVEL,Z,RHCH4,FH2,FARGON,TEMP,PRESS,DEN,XMU,
1
geomalb.f:1083: (continued):
SUBROUTINE ATMSETUP(NLEVEL,Z,RHCH4,FH2,FARGON,TEMP,PRESS,DEN,XMU,
2
Argument #2 (named `z') of `atmsetup' is one type at (2) but is some other type at (1) [info -f g77 M GLOBALS]
Part of the program:
SUBROUTINE ATMSETUP(NLEVEL,Z,RHCH4,FH2,FARGON,TEMP,PRESS,DEN,XMU,
& CH4,H2,XN2,AR,IPRINT)
PARAMETER (NMAX=201)
DIMENSION CH4(1),H2(1),XN2(1),AR(1)
DIMENSION TLINAL(NMAX),DLINAL(NMAX),PLINAL(NMAX)
DIMENSION DEN(1), PRESS(1), TEMP(1), XMU(1)
CALL LINDAL(NLEVEL,Z,TLINAL,DLINAL,PLINAL)
DO J=1,NLEVEL
TEMP(J)= TLINAL(J)
DEN(J)= DLINAL(J)
PRESS(J)=PLINAL(J)
ENDDO
DO 1000 ITS =1,20
CH4(NLEVEL)=PCH4(TEMP(NLEVEL))*RHCH4/PRESS(NLEVEL)
DO 134 J=NLEVEL-1,1,-1
CH4SAT=PCH4(TEMP(J))/PRESS(J)
CH4(J)=AMIN1(CH4SAT,CH4(NLEVEL),CH4(J+1))
134 CONTINUE
DO 20 J=1,NLEVEL
H2(J)=FH2
IF (FARGON .LT. 0.) THEN
AR(J)=(-FARGON-28.0134+25.8554*H2(J)+11.9708*CH4(J))/11.9346
ELSE
IF (FARGON .EQ. 0.) THEN
AR(J)=0.0
ELSE
AR(J)=FARGON
ENDIF
ENDIF
XN2(J)=1.0 - H2(J) - CH4(J) -AR(J)
XMU(J)=28.0134*XN2(J)+2.158*H2(J)+16.0426*CH4(J)+39.948*AR(J)
20 CONTINUE
SUMT=PLINAL(1)*6.02E23/10.
SUMB=SUMT
TLAST=TEMP(NLEVEL)
DO J=2,NLEVEL
DENF=294.1/(XN2(J)*294.1 + CH4(J)*410. + H2(J)*136. + AR(J)*277.8)
DEN(J) = DLINAL(J)*DENF
ADEN=(DEN(J)-DEN(J-1))/ALOG(DEN(J)/DEN(J-1))
SUMT=SUMT+(EFFG(Z(J))*ADEN)*( Z(J-1)-Z(J))*XMU(J)
ADEN=(DLINAL(J)-DLINAL(J-1))/ALOG(DLINAL(J)/DLINAL(J-1))
SUMB=SUMB+(EFFG(Z(J))*ADEN)*( Z(J-1)-Z(J))*28.01340
PRESS(J)=PLINAL(J)*SUMT/SUMB
TEMP(J) =TLINAL(J)*(SUMT/SUMB)*(1./DENF)
ENDDO
30 CONTINUE
DT= ABS(TEMP(NLEVEL)-TLAST)
IF (DT .LT. 0.001) GO TO 1001
1000 CONTINUE
1001 IF (IPRINT .LT. 0) RETURN
WRITE (6,139)RHCH4,FH2,FARGON,DT
DO 135 J=1,NLEVEL-1
WRITE(6,140)J,Z(J),PRESS(J),DEN(J),TEMP(J),
& CH4(J)*PRESS(J)/PCH4(TEMP(J))
& ,CH4(J)*100.,XN2(J)*100.,H2(J)*100.,AR(J)*100.,XMU(J)
& ,(TEMP(J+1)-TEMP(J))/(Z(J+1)-Z(J))
135 CONTINUE
J=NLEVEL
WRITE(6,140)J,Z(J),PRESS(J),DEN(J),TEMP(J),
& CH4(J)*PRESS(J)/PCH4(TEMP(J))
& ,CH4(J)*100.,XN2(J)*100.,H2(J)*100.,AR(J)*100.,XMU(J)
139 FORMAT(///' BACKGROUNG ATMOSPHERE AT LEVELS'/
& ' SURFACE HUMIDITY OF CH4:',F5.3,' H2 MIXING RATIO:',F6.4,
& ' ARGON SETTING:',F8.4/' FINAL CONVERGENCE ON TEMP:',F10.5
& , ' LINDAL ET AL SCALING'/
&' LVL ALTITUDE P(BARS) DEN(CM-3) TEMP RH-CH4'
& , ' %CH4 %N2 %H2 %AR MU DT/DZ' )
140 FORMAT(1X,I3,F8.3,1P2E10.3,0PF7.2,F5.2,2F6.2,2F5.2,4F6.2)
RETURN
END SUBROUTINE
END
And earlier in the program this is how I call the subroutine:
CALL ATMSETUP(NLEVEL,Z,RHCH4,FH2,FARGON,TEMP,PRESS,DEN,XMU,
& CH4,H2,XN2,AR,IPRINT)
You do not provide any type information for your variables and instead rely on implicit typing. Under default implicit typing rules, your variables
Z
andEFFG
are declared as:Both
Z
andEFFG
are real variables of rank 0 (scalar). You reference these variables in the following wayThis is a rank mismatch. You access
Z
as a rank 1 variable andEFFG
as a rank 1 variable, but they were implicitly defined as rank 0 variables.The fix is to properly declare the variables as rank 1 arrays. In the comments you note that switching to gfortran changed the error from a generic "types are different" error into a specific "rank mismatch". This is a much more helpful error message, and others have noted you should be able (once your issues are all fixed) to compile your code with gfortran with no (or only minor) changes over what g77 compiles. It is worthwhile to adopt modern practices and a simple start would be to change
into
Note that this change will not fix your error, as I have not consulted my crystal ball to determine the dimensions of
Z
andEFFG
. All I have done is add explicit types, which should be understood by the ancient g77 compiler (and gfortran in fixed-form). Also note I've assumed default implicit types in determining the explicit types used above, and if they are not what you expected them to be then that is an even better reason to adopt explicit typing.