aer-rc / lnfl Goto Github PK

Mac OSx Monterey 12.4 with gfortran from gcc
$ make -f make_lnfl osxGNUsgl
gave the following errors:

1195 | CALL BUFOUT (LINMRG,RCDHDR(1),LRC) LN09740
| 2
1196 | CALL BUFOUT (LINMRG,VNU3(1),ILNGTH) LN09750
| 1
Error: Type mismatch between actual argument at (1) and actual argument at (2) (REAL(8)/REAL(4)).
../src/lnfl.f:563:29:

563 | CALL CPSTOR_160 (molcpl,npl)
| 1
Error: Actual argument contains too few elements for dummy argument 'molcpl' (38/47) at (1)
../src/lnfl.f:566:29:

566 | CALL CPSTOR_100 (molcpl,npl)
| 1
Error: Actual argument contains too few elements for dummy argument 'molcpl' (38/47) at (1)
../src/lnfl.f:755:26:

755 | CALL BUFOUT (LINFIL,HID(1),NWDLIN) LN04740
| 1
......
1195 | CALL BUFOUT (LINMRG,RCDHDR(1),LRC) LN09740
| 2
Error: Type mismatch between actual argument at (1) and actual argument at (2) (CHARACTER(8)/REAL(4)).
../src/lnfl.f:757:29:

757 | CALL BUFOUT (LINFIL,n_negepp(1),NWDLIN2)
| 1
......
1195 | CALL BUFOUT (LINMRG,RCDHDR(1),LRC) LN09740
| 2
Error: Type mismatch between actual argument at (1) and actual argument at (2) (INTEGER(4)/REAL(4)).
../src/lnfl.f:783:33:

780 | CALL BUFIN (LINMRG,IEOF,RCDHDR(1),LRC) LN04780
| 2
......
783 | CALL BUFIN (LINMRG,IEOF,VNU3(1),ILNGTH) LN04810
| 1
Error: Type mismatch between actual argument at (1) and actual argument at (2) (REAL(8)/REAL(4)).
../src/lnfl.f:827:29:

827 | CALL BUFOUT (LINFIL,VNU3(1),ILNADD)
| 1
......
1195 | CALL BUFOUT (LINMRG,RCDHDR(1),LRC) LN09740
| 2
Error: Type mismatch between actual argument at (1) and actual argument at (2) (REAL(8)/REAL(4)).
make[1]: *** [lnfl.o] Error 1
make: *** [osxGNUsgl] Error 2

Hi,

I'm working on learning LNFL (finally) and would like to see some example TAPE5 files that work with LNFL. In particular, I'm interested in examples (with angular dependence) for UV, VIS, near-IR, IR, far-IR, and MW. Ultimately this will be used for training transmittance regression coefficients in CRTM.

The old TAPE5 files (~10 years old) that we have generate an error mesage forrtl: severe (64): input conversion error, unit 5, file /data/users/bjohnson/CRTM/LNFL/TAPE5 -- so something is wrong with the formatting.

Also, do TAPE5 files for LBLRTM work with LNFL (this might be the problem is that I'm confusing the two).

compile fails on linux with a recent fortran compiler. Using the "-std=legacy" option makes the compiler run through.

I am learning using your LBLRTM. It needs the TAPE3 file, which is created by LNFL. To use LNFL, the ASCII line file parameter database from RTWeb is needed. But http://rtweb.aer.com/line_param_download.php is not accessible. Can I get the file? By the way, is the TAPE5 included in that datase? Your response will be appreciated. Thank you.

Hi,

It seems there is at least one other case of duplicated lines, affecting H2O. This was forwarded to me by my colleague, Romain Allart.

A first case is:
1116984.843045 2.010D-24 9.537E-03.0849.3880 325.34790.62-.032000 54 1 6 0 6 5 0 5 3076244 4
1116984.847100 2.074D-24 9.843E-03.0838.3910 325.34790.36-.026200 57 1 6 1 6 5 0 5 3345334 7
These 2 lines have very similar parameters. The line in bold is not in HITRAN 2020.

He also reported that the 2nd line of the following 3 is likely wrong. But this does not look like a duplication:
1115437.809294 2.980D-24 8.061E-03.0948.4340 173.36580.69-.019000 48 1 4 1 3 3 1 2 4672244 7
1115437.856300 5.954D-25 7.045E-03.0832.3910 326.62550.62-.024589 48 1 6 1 6 5 1 5 3342034 7
1115437.893285 3.311D-26 2.026E-04.0769.4120 382.51690.67-.021191 45 1 5 4 1 4 3 2 2342034 7

Best regards,
Alain Smette

I build LNFL but it failed, the folder named lnfl_v3.2_linux_gnu_sgl.obj hsa been created but is empty. I use gfortan on linux, take make -f make_lnfl linuxGNUsgl.
The gcc and gfortran version is 11 at first, it failed, I changed it to 9 both, it occured like this:

Warning: Possible change of value in conversion from REAL(8) to REAL(4) at (1) [-Wconversion]
../src/lnfl.f:580:30:

579 | DO 10 I = 1, mol_max LN03130
| 2
580 | MINDC(I) = MOLCPL(I)
| 1
Warning: Array reference at (1) out of bounds (47 > 38) in loop beginning at (2)

make[1]: *** [makefile.common:404: lnfl.o] Error 1
make[1]: Leaving directory '/home/zhang/LBLRTM/LNFL/build'
make: *** [makefile.common:213: linuxGNUsgl] Error 2

add TAPE5 for IR and MW

I have so far been unble to compile LNFL (and LBRTM) on my mac monterey and on ubuntu because of type mismatch problems. I am trying to complie with gfortran version 12.3 on my mac and with version 11.3 on ubuntu. The problem is almost certainly associated with the facty that type mismatch issues are now errors as opposed to warnings as indicated here

today.txt

Same issue as AER-RC/LBLRTM#9. I'm posting here as a cross-reference in case other people run into errors like

Error: Actual argument contains too few elements for dummy argument ‘molcpl’ (38/47) at (1)
../src/lnfl.f:566:29:

  566 |             CALL CPSTOR_100 (molcpl,npl)
      |                             1
Error: Actual argument contains too few elements for dummy argument ‘molcpl’ (38/47) at (1)
../src/lnfl.f:3817:18:

 3817 |          STR(I) = STR(I)/(VNU(I)*(1.0-EXP(-BETA0*VNU(I))))               LN15580
      |                  1

with gfortran under Ubuntu 20.04 and later.

Editing the makefile similarly to the recommendation in AER-RC/LBLRTM#9 solves the problem.

I run the LNFL_EXE with TAPE1(downloaded from HITRANonline) in Windows. Some lines were skipped, but I did't do any changes to TAPE1, except to change its name from 5fd07c92.par to TAPE1.
The details are as follows：

LNFL can be completed, but I don't know if the skipping of some lines will affect my subsequent calculation, I like to get your help to find out why lines were skiiped.

Hi,

It seems to me there are still duplicated lines in aer_v3.8.1.
Indeed, the linux command
uniq -d aer_v_3.8.1 | wc -l
suggests that the file has ~ 53,000 duplicated lines.

I could obtain a cleaned version using
mv aer_v_3.8.1 aer_v_3.8.1_orig
uniq aer_v_3.8.1_orig > aer_v_3.8.1

Or do I miss something?

Best regards, Alain Smette

When I tried to complie the lnfl, it said like this.

...
Warning: Possible change of value in conversion from REAL(8) to REAL(4) at (1) [-Wconversion]
../src/lnfl.f:580:30:

  579 |          DO 10 I = 1, mol_max                                            LN03130
      |                                                                        2
  580 |             MINDC(I) = MOLCPL(I)
      |                              1
Warning: Array reference at (1) out of bounds (47 > 38) in loop beginning at (2)
make[1]: *** [makefile.common:404: lnfl.o] Error 1
make: *** [makefile.common:371: mingwGNUsgl] Error 2

I'd appreciate for whom might help.

Thanks.