PySDM icon indicating copy to clipboard operation
PySDM copied to clipboard

Lowe et al. 2019 example mismatches and constant values

Open claresinger opened this issue 2 years ago • 1 comments

This issue is open to document some discrepancies, mismatches, and unknowns about the many constants used in the Lowe et al. 2019 example.

  • [ ] The specification of the aerosol mass compositions and individual component properties is unclear in the paper. There are typos/inconsistencies in the paper itself, but also the paper and the code do not agree on these values. In particular, the inclusion of SOA1 and SOA2 in the Boreal and Nascent aerosols. Some of these disagreements are documented in the aerosol.py vs. aerosol_code.py files. Ideally someone from Dr. Riipinen's group could check our code against the version of their code used for Sam's paper.

  • [ ] Fig 1c and 1d do not match paper. Changing delta_min = 0.1 helps. The value in paper and code is 0.2. Maybe also the temperature is wrong? The temperature is not specified in the paper.

  • [ ] Fig 2a and 2b: The smoke tests for these parcel simulations do not pass for the Boreal and Nascent aerosol cases. They do pass for the Marine case. The parcel results for the Constant and Ovadnevaite surface tensions both fail. This leads me to think it is a problem with the specification of the aerosol composition (see above) and not a problem with the parcel numerics or the surface tension parameterization. However, as described by #1246 the supersaturation at the end of the simulation also shows a mismatch in the Pyrcel example. So there might be the same problem here.

  • [ ] In Fig 2c mean radii of droplets are ~0.5um too small. This is true for the Constant and Ovadnevaite surface tension. Indicates the issues here are not necessarily with the surface partitioning. More likely related to the same problem evident in the smoke tests with the supersaturation and CDNC at maximal altitude.

  • [ ] Fig S2: liquid water path ranges don't match paper supplement. This propagates into mismatches in optical depth and albedo. The paper plot shows LWP ranging from 33 to 28 g/m^2.

claresinger avatar Jan 22 '24 17:01 claresinger

Stale issue message

github-actions[bot] avatar Jul 07 '24 13:07 github-actions[bot]