search

eagles-project / haero

A toolbox for constructing performance portable aerosol packages
Other
3 stars 3 forks source link

Rename process needs mean diameter value for each mode #235

Closed singhbalwinder closed 2 years ago

singhbalwinder commented 3 years ago

In the past, we decided to store only the min and max diameter for a mode in mode related metadata. Rename processes uses nominal diameter (it is not exactly the mean) of a mode to decide whether to trigger inter-mode particle transfer or not. In E3SM, we carry these as meta data. Should we add nominal diameters as meta data to haero as well?

I am calling these values "nominal" as I am not sure how these were decided in the past. They may be empirically derived or they may be a function of some other metadata. I will see if I can find source of these values.

Tagging @jeff-cohere @pbosler @huiwanpnnl @kaizhangpnl

pbosler commented 3 years ago

From the mathematics of the modal method my impression is that "mean diameter" must be diagnosed from the modal PDF --- in this case, I believe it should be the geometric mean, which is dependent on the 0th and 3rd moments of the PDF.

In Haero, you can compute it in 2 steps: first, you need particle volume (3rd moment / 0th). You can get this from our new diagnostic functors. In particular, you need ModeDryParticleVolume. If you need the wet volume, you'll also need to use ModeWetRadius (which requires ModeHygroscopicity's output).

Second, you get the diameter that you need from Mode::mean_particle_diameter_from_volume.

pbosler commented 3 years ago

An example of this is in haero/tests/mode_wet_radius_tests.cpp, lines 115--144 for dry or 115--154 for wet.

singhbalwinder commented 3 years ago

Great! Thanks Pete! I will look into these functions.

pbosler commented 3 years ago

@singhbalwinder, @jeff-cohere The above notes an inconsistency: we have a diagnostic functor for dry particle volume, but wet particle radius. I only did it this way because that's what water uptake needs.

In various places, for both dry and wet particles, both volume and either diameter or radius are used, so we'll frequently need to switch between them with functions like Mode::mean_particle_volume_from_diameter and Mode::mean_particle_diameter_from_volume, which already exist, so the reason for the change would simply be consistency between the diagnostic functors.

jeff-cohere commented 3 years ago

@ashermancinelli , here's some more detail about what we discussed earlier today.

pbosler commented 3 years ago

@ashermancinelli , @singhbalwinder Is mean diameter still an open question for the rename process?

jeff-cohere commented 2 years ago

Closed via #304