Known bugs and issues

Please see our Issue tracker on GitHub for a list of recent bugs and fixes.

Current bug reports

These bug reports (on GitHub) are currently unresolved. We hope to fix these in future releases.

Other issues that you should know about

Bias in tropospheric J-Values when using 47 vertical levels

GEOS-Chem Classic simulations focused on the troposphere can reduce the number of vertical levels from 72 to 47 by coarsening the vertical resolution in the stratosphere and mesosphere. Recent work shows that this vertical remapping does not affect tropospheric transport, but it has a slight effect on tropospheric photolysis rate constants.

To avoid this issue, we recommend one of the following solutions:

  1. Use 72 vertical levels with all GEOS-Chem Classic fullchem simulations,

  2. Use GCHP (either normal grid or stretched grid) fullchem simulations.

GCC 12.2.0 is discontinued in Spack v1.0.0

As of Spack v1.0, spack-packages has been split off into its own separate repository. This change includes the unfortunate deprecation of the GNU Compiler Collection (GCC) version 12.2.0. It appears that only the most recent minor release in each major release is now treated as stable. These deprecations are updated promptly for example, GCC 12.4.0 is already marked as deprecated just 10 days after the release of GCC 12.5.0.

Deprecated GCC versions are no longer listed with the spack info command, so rather than warning users about deprecation, Spack simply fails with an unhelpful error message about not being able to satisfy the request.

For the time being, we recommend that you use Spack release v0.23.1 which still supports GCC 12.2.0 and related libraries. Please see our Build required software with Spack Supplemental Guide for an updated Spack installation workflow.

Discontinuity in GEOS-FP convection at 01 Jun 2020

The convection scheme used to generate archived GEOS-FP meteorology files changed from RAS to Grell-Freitas starting 01 June 2020 with impact on vertical transport. Discussion and analysis of the impact is available at https://github.com/geoschem/geos-chem/issues/1409.

To fix this issue, different GEOS-Chem convection schemes are called based on simulation start time. This ensures comparability in GEOS-Chem runs using GEOS-FP fields generated using the RAS convection scheme and fields generated using Grell-Freitas, but only if the simulation does not cross the 01 June 2020 boundary. We therefore recommend splitting up GEOS-FP runs in time such that a single simulation does not span this date. For example, configure one run to end on 01 June 2020 and then use its output restart to start another run on 01 June 2020. Alternatively consider using MERRA-2 which was entirely generated with RAS, or GEOS-IT which was entirely generated with Grell-Freitas. If you wish to use a GEOS-FP meteorology year different from your simulation year please create a GEOS-Chem GitHub issue for assistance to avoid accidentally using zero convective precipitation flux.

Reaction rate diagnostics may not work as advertized

It has been brought to our attention that the RxnConst and the RxnRates diagnostic collection may not be working as expected. We are currently investigating.

Bugs that have been resolved

These bugs (reported on GitHub) have been resolved.