RTTOV v13 Code Updates and Known Issues

RTTOV v13 Code Updates

Updates and bug fixes to RTTOV v13 are published here. There are separate pages for RTTOV v12 bug fixes and RTTOV v11 bug fixes.

Known bugs and issues

Please report any bugs you find in RTTOV using the NWP SAF Helpdesk. The official release packages for each version don’t contain the bug fixes or updates listed here under that version so you should apply any fixes provided below. Bugs/updates in earlier versions have been fixed/implemented in later versions.

Hide table of known bugs and other updates in RTTOV v13.2

Date Type/Status Description

02/09/2024 Bug - fixed Affects: reading an ASCII rtcoef file at the same time as an HDF5 sccldcoef, scaercoef, and/or pccoef file (RTTOV v13.x, v12.x, v11.x).

Issue: reading may fail with an error such as "Incompatible channels between rtcoef and sccoef files".

Fix: this is quite an obscure bug that is unlikely to affect many users (most people will be reading coefficient files in the same format).

For RTTOV v13.2, download and extract this tarball to the src/ directory, and recompile RTTOV (a clean compilation is not necessary).

For earlier RTTOV versions please contact the NWP SAF helpdesk.

05/07/2024 Bug - fixed Affects: IR emissivity atlases: UWIRemis, CAMEL single-year, and CAMEL climatology atlases (all RTTOV versions in which they are implemented).

Issue: there may be intermittent memory-related failures with some compilers when initialising the atlases for use with a single instrument when the covariance data are not being used, due to a missing initialisation of a variable that is used as a dimension for some automatic arrays. This can lead to the code requesting arbitrary amounts of memory.

Fix: for v13.x download and extract download this tarball to the src/emis_atlas/ directory, and recompile RTTOV (a clean compilation is not necessary).

For earlier RTTOV versions please contact the NWP SAF helpdesk.

29/05/2024 Bug - fixed Affects: RTTOV-SCATT simulations (all versions of RTTOV back to v9 at least).

Issue: in cases where the total asymmetry parameter of all hydrometeors combined is negative RTTOV-SCATT was not computing the total value correctly. The impact of the bug in most cases is very small, but for certain profiles can be as large as several tenths of a Kelvin (particularly for lower frequency channels, for example, at ~30 GHz and below).

Fix: for v13.2 download and extract this tarball to the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary).

For v13.0/v13.1 use this tarball instead.

NB these tarballs include the v13.1 RTTOV-SCATT bug fix dated 11/02/2022.

For earlier RTTOV versions please contact the NWP SAF helpdesk.

21/05/2024 Bug - fixed Affects: RTTOV-SCATT simulations with GPM GMI, FY-3F MWRI-2, and FY-3G MWRI-RM optical depth coefficients based on measured SRFs (all RTTOV versions compatible with the above SRF-based coefficients, at least v13.x, v12.x).

Issue: pairs of consecutive imager channels at the same nominal frequency with different polarisations should use the same table of optical properties from the hydrotable. When measured SRFs are used, channel central frequencies can differ, and for the above instruments, they differ by more than the threshold used in the code to identify such pairs of channels. This leads to the wrong optical properties being used for some channels and the impact on radiances can be very large.

Fix: for RTTOV v13.x download this file to the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary).

For fixes for v12.x and earlier please contact the NWP SAF help desk.

15/12/2023 Bug - fixed Affects: RTTOV test suite when compiling with HDF5 library v1.12.3 or later.

Issue: RTTOV attempts to re-open HDF5 library without first closing it which only causes errors with the latest versions of the HDF5 library.

Fix: for RTTOV v13.2 download this file to the src/coef_io/ directory, and recompile RTTOV (a clean compilation is not necessary).

For fixes for v13.1 and v13.0 please contact the NWP SAF help desk.

10/12/2023 Bug - fixed Affects: MFASIS-NN (v13.2 only) and MFASIS-LUT (v13.x, v12.x)

Issue: there is a theoretical possibility of numerical errors resulting from division by zero, but in practice this is unlikely to occur. No issues have been observed in any simulations by members of the development team, nor reported by any users.

Fix: for v13.2 download and extract download this tarball to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).

This fix must be applied after the MFASIS-LUT bug fix below dated 02/08/2023.

Note that for MFASIS-NN this results in very small changes in simulated reflectances (in the 6th or greater significant figure).

For fixes for v13.1 and v13.0 please contact the NWP SAF help desk.

02/08/2023 Bug - fixed Affects: MFASIS LUT simulations for EPIC, AMI and ABI with the new v13.1 LUT files (v13.2, v13.1).

Issue: there is an indexing error which means the final supported channel in each case (ch 6 for EPIC and AMI, ch 5 for ABI) may not be simulated correctly. There is a code error and also an error in the AMI and ABI LUT files: updated files are available and should be downloaded and used with the code patch below.

Fix: for v13.2 download and extract download this tarball to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).

For v13.1, use this tarball instead.

16/05/2023 Bug - fixed Affects: quality flags for MFASIS-NN simulations. Output radiances/reflectances are not affected (v13.2).

Issue: rounding errors can cause the quality flag indicating the effective diameter bounds have been exceeded to be set when this is not the case.

Fix: download this file to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).

30/03/2023 Bug - fixed Affects: rttov_coef_download.sh script for downloading RTTOV v13 coefficient files.

Issue: there is a typo that means the script does not download MFASIS-NN coefficient files.

Fix: here is a fixed version of the script.

01/12/2020 Bug - outstanding Affects: HTFRTC simulations (v13.x, v12.x)

Issue: there is a bug in HTFRTC which results in zero radiances for very dry profiles.

Fix: there is no fix available for this due to resourcing issues. HTFRTC coefficient files for RTTOV v13 will not be made available until the bug is fixed. Note that the v12 files are not compatible with v13. As such it is not currently possible to run HTFRTC in RTTOV v13. It is not recommended to use HTFRTC in any RTTOV version, and it will be removed in RTTOV v14.

Date	Type/Status	Description
02/09/2024	Bug - fixed	Affects: reading an ASCII rtcoef file at the same time as an HDF5 sccldcoef, scaercoef, and/or pccoef file (RTTOV v13.x, v12.x, v11.x). Issue: reading may fail with an error such as "Incompatible channels between rtcoef and sccoef files". Fix: this is quite an obscure bug that is unlikely to affect many users (most people will be reading coefficient files in the same format). For RTTOV v13.2, download and extract this tarball to the src/ directory, and recompile RTTOV (a clean compilation is not necessary). For earlier RTTOV versions please contact the NWP SAF helpdesk.
05/07/2024	Bug - fixed	Affects: IR emissivity atlases: UWIRemis, CAMEL single-year, and CAMEL climatology atlases (all RTTOV versions in which they are implemented). Issue: there may be intermittent memory-related failures with some compilers when initialising the atlases for use with a single instrument when the covariance data are not being used, due to a missing initialisation of a variable that is used as a dimension for some automatic arrays. This can lead to the code requesting arbitrary amounts of memory. Fix: for v13.x download and extract download this tarball to the src/emis_atlas/ directory, and recompile RTTOV (a clean compilation is not necessary). For earlier RTTOV versions please contact the NWP SAF helpdesk.
29/05/2024	Bug - fixed	Affects: RTTOV-SCATT simulations (all versions of RTTOV back to v9 at least). Issue: in cases where the total asymmetry parameter of all hydrometeors combined is negative RTTOV-SCATT was not computing the total value correctly. The impact of the bug in most cases is very small, but for certain profiles can be as large as several tenths of a Kelvin (particularly for lower frequency channels, for example, at ~30 GHz and below). Fix: for v13.2 download and extract this tarball to the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary). For v13.0/v13.1 use this tarball instead. NB these tarballs include the v13.1 RTTOV-SCATT bug fix dated 11/02/2022. For earlier RTTOV versions please contact the NWP SAF helpdesk.
21/05/2024	Bug - fixed	Affects: RTTOV-SCATT simulations with GPM GMI, FY-3F MWRI-2, and FY-3G MWRI-RM optical depth coefficients based on measured SRFs (all RTTOV versions compatible with the above SRF-based coefficients, at least v13.x, v12.x). Issue: pairs of consecutive imager channels at the same nominal frequency with different polarisations should use the same table of optical properties from the hydrotable. When measured SRFs are used, channel central frequencies can differ, and for the above instruments, they differ by more than the threshold used in the code to identify such pairs of channels. This leads to the wrong optical properties being used for some channels and the impact on radiances can be very large. Fix: for RTTOV v13.x download this file to the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary). For fixes for v12.x and earlier please contact the NWP SAF help desk.
15/12/2023	Bug - fixed	Affects: RTTOV test suite when compiling with HDF5 library v1.12.3 or later. Issue: RTTOV attempts to re-open HDF5 library without first closing it which only causes errors with the latest versions of the HDF5 library. Fix: for RTTOV v13.2 download this file to the src/coef_io/ directory, and recompile RTTOV (a clean compilation is not necessary). For fixes for v13.1 and v13.0 please contact the NWP SAF help desk.
10/12/2023	Bug - fixed	Affects: MFASIS-NN (v13.2 only) and MFASIS-LUT (v13.x, v12.x) Issue: there is a theoretical possibility of numerical errors resulting from division by zero, but in practice this is unlikely to occur. No issues have been observed in any simulations by members of the development team, nor reported by any users. Fix: for v13.2 download and extract download this tarball to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). This fix must be applied after the MFASIS-LUT bug fix below dated 02/08/2023. Note that for MFASIS-NN this results in very small changes in simulated reflectances (in the 6th or greater significant figure). For fixes for v13.1 and v13.0 please contact the NWP SAF help desk.
02/08/2023	Bug - fixed	Affects: MFASIS LUT simulations for EPIC, AMI and ABI with the new v13.1 LUT files (v13.2, v13.1). Issue: there is an indexing error which means the final supported channel in each case (ch 6 for EPIC and AMI, ch 5 for ABI) may not be simulated correctly. There is a code error and also an error in the AMI and ABI LUT files: updated files are available and should be downloaded and used with the code patch below. Fix: for v13.2 download and extract download this tarball to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). For v13.1, use this tarball instead.
16/05/2023	Bug - fixed	Affects: quality flags for MFASIS-NN simulations. Output radiances/reflectances are not affected (v13.2). Issue: rounding errors can cause the quality flag indicating the effective diameter bounds have been exceeded to be set when this is not the case. Fix: download this file to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).
30/03/2023	Bug - fixed	Affects: rttov_coef_download.sh script for downloading RTTOV v13 coefficient files. Issue: there is a typo that means the script does not download MFASIS-NN coefficient files. Fix: here is a fixed version of the script.
01/12/2020	Bug - outstanding	Affects: HTFRTC simulations (v13.x, v12.x) Issue: there is a bug in HTFRTC which results in zero radiances for very dry profiles. Fix: there is no fix available for this due to resourcing issues. HTFRTC coefficient files for RTTOV v13 will not be made available until the bug is fixed. Note that the v12 files are not compatible with v13. As such it is not currently possible to run HTFRTC in RTTOV v13. It is not recommended to use HTFRTC in any RTTOV version, and it will be removed in RTTOV v14.

Hide table of known bugs and other updates in RTTOV v13.1. Bugs listed above may also affect v13.1.

Date Type/Status Description

19/07/2023 Bug - fixed Affects: BRDFs returned from the BRDF atlas for channels at wavelengths in the range 0.4-0.4002 microns (v13.1, v13.0, v12.x, v11.3).

Issue: an error in the snow BRDF PC eigenvectors for the shortest wavelength causes spurious values to be returned for snow-affected locations. Almost no instruments are affected: one channel of GOME2 is the only known case among sensors supported by RTTOV.

Fix: for RTTOV v13.1 and v13.0 download this file to the src/brdf_atlas/ directory, and recompile RTTOV (a clean compilation is not necessary). If you require a fix for RTTOV v12.x or v11.3 please contact the NWP SAF help desk.

25/11/2022 Compiler issue - mitigated. Affects: compilation of RTTOV using recent versions of gfortran, at least v10, v11 (RTTOV v13.1, v13.0, v12.x, v11.x).

Issue: compilation seems to hang when it reaches the src/other/rttov_bpr_init.F90 subroutine. Compilation will eventually succeed, but it takes a very long time. This is a problem with these versions of gfortran, not the RTTOV code.

Fix: for RTTOV v13.1, v13.0 and v12.x download this file to the src/other/ directory, and recompile RTTOV (a clean compilation is not necessary).

For RTTOV v11.x use this file instead.

26/10/2022 Bug - fixed Affects: volcanic ash particle types ("VOLA" and "VAPO") in OPAC scaercoef aerosol optical property files when supplying aerosol concentration inputs in units of kg/kg. Other particle types and number density input units are unaffected. (RTTOV v13.x, v12.x).

Issue: the factors used by RTTOV to convert volcanic ash aerosol concentrations from kg/kg to number density were wrong by a factor of 10^6. This resulted in aerosol simulations assuming lower concentrations than those input.

Fix: all OPAC aerosol optical property files have been updated with the corrected factors. These are available from the coefficients download page.

29/06/2022 Bug - fixed Affects: VIS/IR cloud simulations using the Baran scheme for ice cloud optical properties (RTTOV v13.x, v12.x, v11.x, but note that it particularly affects the far-IR, especially channels around 100-500 cm-1, and the impact is only significant for the DOM solver).

Issue:
1. it is possible for the Baran parameterisation to yield negative values for bpr (the backscattered fraction of radiation) and values of the asymmetry parameter larger than one. This occurs for warm clouds with high ice water content and as such is relatively rare. These unphysical values also primarily occur in the far-IR, around 100-500 cm-1. The bpr values are used by the Chou-scaling solver (although impacts due to this issue are extremely small). The asymmetry parameter is used by the DOM solver, and these unphysical values can cause numerical failures for affected profiles.

2. for DOM simulations the Baran parameterisation generates a phase function based on the asymmetry parameter. This is interpolated onto a Gaussian quadrature in order to compute the Legendre coefficients. This process has previously used a spline interpolation, but for asymmetry parameter values close to one (which particularly occur in the far-IR), this can yield negative lobes in the interpolated phase function. In the fix below, the spline interpolation has been replaced with a simple linear interpolation which eliminates the unphysical interpolated phase functions, but changes all results slightly when using the DOM solver. The impact in the visible/near-IR is very small (e.g. of order 1.E-4 in reflectance). In the IR the impact can be up to order 0.1 K, but in most cases is an order of magnitude smaller than this.

Fix: download this tarball and extract within the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). See above for a description of the impact of this fix. This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for an earlier version.

24/05/2022 Bug - to be fixed in next release Affects: MW simulations in RTTOV and RTTOV-SCATT when any channel has its "validity flag" set to zero (indicating the channel should not be simulated) in the rtcoef file. (RTTOV v13.x, v12.x, v11.x).

Issue: for MW simulations, checks are not in place to avoid calculations for deactivated channels which can lead to failures or NaNs in the output. NB No current or recent MW rtcoef files have any validity flags set to zero so this problem does not occur unless you run with certain RTTOV v10-era MW coefficient files or if you edit the validity flags in the coefficient files.

Fix: this issue will be fixed in the code for RTTOV v13.2 and future versions. However it is simple to work around: either set all the validity flags in the coefficient file to 1 and ignore outputs in the previously deactivated channels, or do not simulate the affected channels by omitting them when populating the rttov_chanprof structure.

24/05/2022 Bug - fixed Affects: RTTOV-SCATT AD/K simulations (RTTOV v13.x and earlier versions)

Issue: in rare cases the adjoint code is numerically unstable.

Fix: download this tarball and extract within the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary). The impact of the fix is generally well below 1 mK on direct model radiances. This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for an earlier version.

The fix changes test reference output for the test_rttovscatt.sh script. Download this updated test reference data and extract in rttov_test/.

11/02/2022 Bugs - fixed Affects: FASTEM-3 emissivities for all sensors, and FASTEM-4/5 emissivities for Stokes 3/4 components for polarimetric sensors only (all versions of RTTOV from v13.1 back to v9 at least).

Issues:
1. FASTEM-3: the definition of azimuth angle was based on satellite azimuthal view angle instead of the documented RTTOV definition of satellite azimuth. The relative wind direction computed was therefore wrong by 180 degrees. This significantly affects the azimuthal dependence of V-pol emissivities (previously this was very different to other FASTEM versions, with the fix below it is much more consistent). It has a smaller impact on H-pol emissivities. For Stokes 3 components, emissivities have a phase shift of 180 degrees. There is also a smaller impact on Stokes 4 components.

2. FASTEM-4/5: the definition of relative wind direction was the negative of what it should be. This has no impact on V- and H-pol emissivities, but the emissivities for Stokes 3 and 4 components have the wrong sign.

Fix: these bugs have been fixed. Note that the FASTEM-3 fix will be implemented on a switch in RTTOV v13.2 to support any users who are happily using the existing FASTEM-3 emissivities. Experiments with the fixed code indicate that the FASTEM-3 treatment of Stokes 3/4 channels fits observations better than that for FASTEM-4/5. Previously FASTEM-6 has not supported Stokes 3/4 channels: the update below, in addition to fixing the above bugs, implements the FASTEM-3 treatment of Stokes 3/4 channels in FASTEM-6. With this update, FASTEM-6 is the recommended emissivity model for all sensor channels below ~200GHz.

To fix the bugs (and enable FASTEM-6 support for polarimetric sensors), download and extract this tarball to src/main/ and recompile RTTOV. A clean build is not required.

These bug fixes affect reference data for the test_fwd.sh and test_rttov13.sh test scripts: updated test reference data are available and should be extracted in the rttov_test/ directory. These reference data are only for RTTOV v13.1.

This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for RTTOV v12.x.

11/02/2022 Bug - fixed Affects: RTTOV-SCATT AD/K simulations involving the new ice polarisation treatment, i.e. MW imagers only (v13.1, v13.0).

Issue: there is an AD coding error. The impact is extremely small (for example, it has not been causing convergence problems in the IFS where it is used operationally) because it relates to the lower extinction threshold (1E-10). It only applies to imagers when the new ice polarisation feature is active.

Fix: see the RTTOV v13.2 bug fix for RTTOV-SCATT dated 29/05/2024. The tarball for v13.0/v13.1 includes the fix for this bug.

Date	Type/Status	Description
19/07/2023	Bug - fixed	Affects: BRDFs returned from the BRDF atlas for channels at wavelengths in the range 0.4-0.4002 microns (v13.1, v13.0, v12.x, v11.3). Issue: an error in the snow BRDF PC eigenvectors for the shortest wavelength causes spurious values to be returned for snow-affected locations. Almost no instruments are affected: one channel of GOME2 is the only known case among sensors supported by RTTOV. Fix: for RTTOV v13.1 and v13.0 download this file to the src/brdf_atlas/ directory, and recompile RTTOV (a clean compilation is not necessary). If you require a fix for RTTOV v12.x or v11.3 please contact the NWP SAF help desk.
25/11/2022	Compiler issue - mitigated.	Affects: compilation of RTTOV using recent versions of gfortran, at least v10, v11 (RTTOV v13.1, v13.0, v12.x, v11.x). Issue: compilation seems to hang when it reaches the src/other/rttov_bpr_init.F90 subroutine. Compilation will eventually succeed, but it takes a very long time. This is a problem with these versions of gfortran, not the RTTOV code. Fix: for RTTOV v13.1, v13.0 and v12.x download this file to the src/other/ directory, and recompile RTTOV (a clean compilation is not necessary). For RTTOV v11.x use this file instead.
26/10/2022	Bug - fixed	Affects: volcanic ash particle types ("VOLA" and "VAPO") in OPAC scaercoef aerosol optical property files when supplying aerosol concentration inputs in units of kg/kg. Other particle types and number density input units are unaffected. (RTTOV v13.x, v12.x). Issue: the factors used by RTTOV to convert volcanic ash aerosol concentrations from kg/kg to number density were wrong by a factor of 10^6. This resulted in aerosol simulations assuming lower concentrations than those input. Fix: all OPAC aerosol optical property files have been updated with the corrected factors. These are available from the coefficients download page.
29/06/2022	Bug - fixed	Affects: VIS/IR cloud simulations using the Baran scheme for ice cloud optical properties (RTTOV v13.x, v12.x, v11.x, but note that it particularly affects the far-IR, especially channels around 100-500 cm-1, and the impact is only significant for the DOM solver). Issue: 1. it is possible for the Baran parameterisation to yield negative values for bpr (the backscattered fraction of radiation) and values of the asymmetry parameter larger than one. This occurs for warm clouds with high ice water content and as such is relatively rare. These unphysical values also primarily occur in the far-IR, around 100-500 cm-1. The bpr values are used by the Chou-scaling solver (although impacts due to this issue are extremely small). The asymmetry parameter is used by the DOM solver, and these unphysical values can cause numerical failures for affected profiles. 2. for DOM simulations the Baran parameterisation generates a phase function based on the asymmetry parameter. This is interpolated onto a Gaussian quadrature in order to compute the Legendre coefficients. This process has previously used a spline interpolation, but for asymmetry parameter values close to one (which particularly occur in the far-IR), this can yield negative lobes in the interpolated phase function. In the fix below, the spline interpolation has been replaced with a simple linear interpolation which eliminates the unphysical interpolated phase functions, but changes all results slightly when using the DOM solver. The impact in the visible/near-IR is very small (e.g. of order 1.E-4 in reflectance). In the IR the impact can be up to order 0.1 K, but in most cases is an order of magnitude smaller than this. Fix: download this tarball and extract within the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). See above for a description of the impact of this fix. This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for an earlier version.
24/05/2022	Bug - to be fixed in next release	Affects: MW simulations in RTTOV and RTTOV-SCATT when any channel has its "validity flag" set to zero (indicating the channel should not be simulated) in the rtcoef file. (RTTOV v13.x, v12.x, v11.x). Issue: for MW simulations, checks are not in place to avoid calculations for deactivated channels which can lead to failures or NaNs in the output. NB No current or recent MW rtcoef files have any validity flags set to zero so this problem does not occur unless you run with certain RTTOV v10-era MW coefficient files or if you edit the validity flags in the coefficient files. Fix: this issue will be fixed in the code for RTTOV v13.2 and future versions. However it is simple to work around: either set all the validity flags in the coefficient file to 1 and ignore outputs in the previously deactivated channels, or do not simulate the affected channels by omitting them when populating the rttov_chanprof structure.
24/05/2022	Bug - fixed	Affects: RTTOV-SCATT AD/K simulations (RTTOV v13.x and earlier versions) Issue: in rare cases the adjoint code is numerically unstable. Fix: download this tarball and extract within the src/mw_scatt/ directory, and recompile RTTOV (a clean compilation is not necessary). The impact of the fix is generally well below 1 mK on direct model radiances. This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for an earlier version. The fix changes test reference output for the test_rttovscatt.sh script. Download this updated test reference data and extract in rttov_test/.
11/02/2022	Bugs - fixed	Affects: FASTEM-3 emissivities for all sensors, and FASTEM-4/5 emissivities for Stokes 3/4 components for polarimetric sensors only (all versions of RTTOV from v13.1 back to v9 at least). Issues: 1. FASTEM-3: the definition of azimuth angle was based on satellite azimuthal view angle instead of the documented RTTOV definition of satellite azimuth. The relative wind direction computed was therefore wrong by 180 degrees. This significantly affects the azimuthal dependence of V-pol emissivities (previously this was very different to other FASTEM versions, with the fix below it is much more consistent). It has a smaller impact on H-pol emissivities. For Stokes 3 components, emissivities have a phase shift of 180 degrees. There is also a smaller impact on Stokes 4 components. 2. FASTEM-4/5: the definition of relative wind direction was the negative of what it should be. This has no impact on V- and H-pol emissivities, but the emissivities for Stokes 3 and 4 components have the wrong sign. Fix: these bugs have been fixed. Note that the FASTEM-3 fix will be implemented on a switch in RTTOV v13.2 to support any users who are happily using the existing FASTEM-3 emissivities. Experiments with the fixed code indicate that the FASTEM-3 treatment of Stokes 3/4 channels fits observations better than that for FASTEM-4/5. Previously FASTEM-6 has not supported Stokes 3/4 channels: the update below, in addition to fixing the above bugs, implements the FASTEM-3 treatment of Stokes 3/4 channels in FASTEM-6. With this update, FASTEM-6 is the recommended emissivity model for all sensor channels below ~200GHz. To fix the bugs (and enable FASTEM-6 support for polarimetric sensors), download and extract this tarball to src/main/ and recompile RTTOV. A clean build is not required. These bug fixes affect reference data for the test_fwd.sh and test_rttov13.sh test scripts: updated test reference data are available and should be extracted in the rttov_test/ directory. These reference data are only for RTTOV v13.1. This fix applies to RTTOV v13.x. Please contact the help desk if you require a fix for RTTOV v12.x.
11/02/2022	Bug - fixed	Affects: RTTOV-SCATT AD/K simulations involving the new ice polarisation treatment, i.e. MW imagers only (v13.1, v13.0). Issue: there is an AD coding error. The impact is extremely small (for example, it has not been causing convergence problems in the IFS where it is used operationally) because it relates to the lower extinction threshold (1E-10). It only applies to imagers when the new ice polarisation feature is active. Fix: see the RTTOV v13.2 bug fix for RTTOV-SCATT dated 29/05/2024. The tarball for v13.0/v13.1 includes the fix for this bug.

Hide table of known bugs and other updates in RTTOV v13.0. Bugs listed above may also affect v13.0.

Date Type/Status Description

31/08/2021 Bug - fixed Affects: VIS/IR cloud simulations using the Ou and Liou ice effective diameter parameterisation (icede_param = 1) with the Baum ice properties (v13.0 only).

Issue: the Ou and Liou parameterisation is valid for temperatures in the range -60 to -20 deg C. The lower bound is incorrectly applied meaning that for temperatures below ~205 K the calculated effective diameters are set to a minimum of 10 microns instead of 22 microns. To give an estimate of the impact, the fixed code has been compared to the original for the eight SEVIRI IR channels over the 137L NWP SAF profile dataset (25000 profiles). The bias across all channels and profiles is ~0.2 K, standard deviation ~0.8 K, and maximum difference ~26 K.

Fix: download and extract this tarball in the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).

31/08/2021 Bug - fixed Affects: MFASIS TL/AD/K models for profiles with mixed phase cloud (v13.0 only).

Issue: a TL/AD/K coding error yields incorrect results for some profiles with layers containing both liquid and ice cloud.

Fix: download and extract this tarball in the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary).

14/06/2021 Bug - fixed Affects: FASTEM 3-6 10m wind u and v component TL/AD/K (all versions of RTTOV from v13.0 back to v9 at least).

Issue: there is a TL/AD/K bug in the FASTEM code related to the wind azimuth direction calculation. This affects all versions of FASTEM that account for azimuthal variation (FASTEM-3 and later). Wind u/v adjoints and Jacobians are incorrect when the input wind u and v have opposite signs. The direct model is not affected.

Fix: download and extract this tarball in the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). This fix is for RTTOV v13.0 only. See the RTTOV v12 bug fix table for v12 fixes.

This bug fix affects reference data for the test_rttov13.sh and test_multi_instrument.sh test scripts: updated test reference data are available and should be extracted in the rttov_test/ directory. These reference data are only for RTTOV v13.0.

11/06/2021 Bug - fixed Affects: NLTE bias correction (RTTOV v13.0, v12.x).

Issue: the temperature predictors calculated for the NLTE bias correction did not average temperature values over the correct set of levels leading to errors in the computed bias correction. An indication of the impact of the bug is given in this plot which shows IASI BT differences for the fixed code minus the original code for the NLTE-affected band for an 83 diverse profile set simulated for satellite zenith angles between 0-60 degrees and solar zenith angles 0-80 degrees, each in steps of 10 degrees, with all combinations of angles tested.

Fix: download and extract this tarball in the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). This fix can be applied to RTTOV v13.0, v12.3, v12.2 and v12.1.

20/05/2021 Bug - fixed Affects: PC-RTTOV simulations with aerosols when the interpolator is not used. (RTTOV v13.0, v12.3, v12.2).

Issue: two array indices were the wrong way around which causes run-time failures.

Fix: download this source file to the src/main/ directory, and recompile RTTOV (a clean compilation is not necessary). This fix can be applied to RTTOV v13.0, v12.3 and v12.2.

13/04/2021 Bug - fixed Affects: RTTOV-SCATT hydrotable generation when reading the ARTS database. The problem was reported for Ubuntu 20.04 LTS and gcc/gfortran 9.3.0, but the bug does not affect other compilers/platforms tested by the RTTOV development team. The bug does not affect the hydrotables available for download on the coefficients download web page, nor does it affect any RTTOV-SCATT simulations. (RTTOV v13.0 only).

Issue: the code which ingests ARTS database reads more elements than are in the file which is assumed to trigger a linefeed, but does not do so for certain compilers/platforms (see above) in which case the output hydrotable may contain erroneous optical properties.

Fix: download this source file to the src/mw_scatt_coef/ directory and recompile RTTOV (a clean compilation is not necessary).

01/02/2021 Bug - fixed Affects: rttov_coef_download.sh script for downloading RTTOV v13 coefficient files.

Issue: if you choose the option not to continue previously interrupted downloads the script does not work properly. It works fine if you select "y" to this question.

Fix: here is a fixed version of the script.

17/12/2020 Bug - fixed Affects: RTTOV-SCATT hydrotable generation code when different permittivity models are specified for different hydrometeors. This is an unlikely usage scenario and the bug does not affect the hydrotables available for download on the coefficients download web page, nor does it affect any RTTOV-SCATT simulations. (RTTOV v13.0 only).

Issue: incorrect array indexes mean that the water/ice permittivity models for the previous hydrometeor are used for the current hydrometeor instead of those specified by the user.

Fix: download this source file to the src/mw_scatt_coef/ directory and recompile RTTOV (a clean compilation is not necessary).