MORSE Bugs

Bugs recorded for MORSE v3 (12SEP05 onwards) - Bugs#1:20 for earlier versions are here

Bug#32
Reported: 3 Jul 09 by Luis Millan-Valle (U.Oxford)
Symptom: Retrieving pT from altitudes above lowest tangent point in L1C data results in Fatal error message
'F-SWPRTV: Singular matrix#3'
Cause: various problems in handling of aerosol Jacobians (RFM Bug#74 and MORSE not setting forward model altitudes below lowest retrieval level, subroutines PTBATM, VECHYD, LEVFMZ.
Fixed: v09JUL09

Bug#31
Reported: 9 Jul 08 by David Moore (U.Leicester)
Symptom: Inconsistent units in aerosol output files
Cause: various problems in handling of aerosol and continuum profiles in subroutines L2PHGT,MSEATM,OUTATM,SWPATM, although shouldn't affect results of other retrieved species
Fixed: v17SEP08

Bug#30
Reported: 14 Jan 08 by Joanne Walker
Symptom: Fails to load (non-LTE) vibrational temperature profiles from files in *CLI section of driver table (a priori VT-KT defaults to 0 K difference)
Cause: incorrect index in loading array TVBNTE in subroutine vibnte.for and remapping in l2phgt.for
Fixed: v14JAN08

Bug#29
Reported: 13 Sep 07 by Joanne Walker
Symptom: Stops with
F-SWPAPR: SA is Singular statement executed
message printed to terminal when using a prescribed A Priori Covariance matrix (in the driver table
*ACV section) of incompatible length with the retrieval state vector.
Cause: Subroutine drvacv.for checks the dimension of the A Priori covariance read from the file for compatibility with the retrieval state vector, detects the error, writes an appropriate error message internally, but fails to exit with fail flag set true so the error is ignored until subroutine swpapr realises that it cannot invert the covariance matrix.
Fixed: v13SEP07

Bug#28
Reported: 16 Aug 07 by Joanne Walker
Symptom: Loading incorrect vibrational temperature profile from ".atm file" when actually using ".rtv file" as input
Cause: .rtv files contain SD profiles, eg 'CO(1)(2)_SD' which was intepreted as the VT profile 'CO(1)(2)' - now fixed to produce a fatal error message since you should not use .rtv files as inputs (always use the equivalent .atm file which contains no SD profile but does contain the full atmospheric profile outside just the retrieval levels)
Fixed: v16AUG07

Bug#27
Reported: 07 Aug 07 by Joanne Walker
Symptom: Segmentation fault during vibrational temperature retrieval when no vib tem data is supplied
Cause: Internally the NTE flag is set when vib tem profiles are read in, but not when vib tem retrieval is specified
Fixed: v07AUG07

Bug#26
Reported: 15 May 07 by Joanne Walker
Symptom: Program hangs while reading *CLI section of driver table.
Cause: Missing '*END' record at end of an .atm file causes infinite DO ... WHILE loop since end of record is not handled by ERR=900 branch in READ statement.
Fixed: v15MAY07

Bug#25
Reported: 22 Feb 07 by Chiara Piccolo
Symptom: Program stops with error message
F-NXTREC: Extra characters in input file beyond C*80: ...
when reading .lst file (for which records up to C*120 are allowed)
Cause: Generic subroutine opnfil.for, which opens the .lst file, calls generic subroutine nxtrec.for which performs a check on records being LE C*80. Now changed so that opnfil.for ignores this particular error message from nxtrec.for
Fixed: v22FEB07

Bug#24
Reported: 26 Apr 06 by Chiara Piccolo
Symptom: Top profile level in *CLI section not overwritten when remainder of profile is replaced by subsequent file.
Cause: Error in subroutine drvcli.for in determining no. of profile levels that should be replaced.
Fixed: v26APR06

Bug#23
Reported: 24 Feb 06 by Chiara Piccolo
Symptom: Isotopic profiles incorrectly set up if not explicilty specified in .atm files (eg *O3 doesn't give correct isotope profiles, but *O3(1), *O3(2) etc do)
Cause: When setting default isotopic profiles each successive profile was being reduced by factor 1E-6 (repeated ppmv to ppv conversion)
Fixed: v01MAR06

Bug#22
Reported: 17 Jan 06 by Joanne Walker
Symptom: Array bound error or other unpredictable behaviour if microwindow with large number of irregular grid points used (eg N2O50101)
Cause: No check if no. irregular grid points (NGRD) for microwindow exceeds available array space (MAXGRD) in rfmsiz.inc (normally set to 5000).
Fixed: v01JAN06

Bug#21
Reported: 31 Oct 05 by Chiara Piccolo
Symptom: A Priori output represents a priori at start of last microwindow rather than first
Cause: When subroutine mseout overwriting a priori output each microwindow rather than only on the first microwindow.
Fixed: v01NOV05